docs/unix-phil

annotate unix-phil.ms @ 49:2f188a5dc0dc

moved style to separate document
author meillo@marmaro.de
date Mon, 12 Apr 2010 14:49:13 +0200
parents 40caeb9e9b25
children bb5901b17889
rev   line source
meillo@49 1 .so style
meillo@42 2
meillo@0 3 .TL
meillo@42 4 .ps +4
meillo@6 5 Why the Unix Philosophy still matters
meillo@0 6 .AU
meillo@0 7 markus schnalke <meillo@marmaro.de>
meillo@0 8 .AB
meillo@1 9 .ti \n(.iu
meillo@39 10 This paper explains the importance of the Unix Philosophy for software design.
meillo@0 11 Today, few software designers are aware of these concepts,
meillo@39 12 and thus a lot of modern software is more limited than necessary
meillo@39 13 and makes less use of software leverage than possible.
meillo@38 14 Knowing and following the guidelines of the Unix Philosophy makes software more valuable.
meillo@0 15 .AE
meillo@0 16
meillo@2 17 .FS
meillo@2 18 .ps -1
meillo@39 19 This paper was prepared for the ``Software Analysis'' seminar at University Ulm.
meillo@47 20 Mentor was professor Franz Schweiggert.
meillo@47 21 Handed in on 2010-04-12.
meillo@39 22 You may retrieve this document from
meillo@39 23 .CW \s-1http://marmaro.de/docs \ .
meillo@2 24 .FE
meillo@2 25
meillo@48 26 .H 1 Introduction
meillo@0 27 .LP
meillo@40 28 The Unix Philosophy is the essence of how the Unix operating system,
meillo@40 29 especially its toolchest, was designed.
meillo@40 30 It is no limited set of fixed rules,
meillo@40 31 but a loose set of guidelines which tell how to write software that
meillo@40 32 suites well into Unix.
meillo@40 33 Actually, the Unix Philosophy describes what is common to typical Unix software.
meillo@40 34 The Wikipedia has an accurate definition:
meillo@40 35 .[
meillo@44 36 wikipedia
meillo@44 37 unix philosophy
meillo@40 38 .]
meillo@40 39 .QP
meillo@40 40 The \fIUnix philosophy\fP is a set of cultural norms and philosophical
meillo@40 41 approaches to developing software based on the experience of leading
meillo@40 42 developers of the Unix operating system.
meillo@1 43 .PP
meillo@40 44 As there is no single definition of the Unix Philosophy,
meillo@40 45 several people have stated their view on what it comprises.
meillo@1 46 Best known are:
meillo@1 47 .IP \(bu
meillo@1 48 Doug McIlroy's summary: ``Write programs that do one thing and do it well.''
meillo@1 49 .[
meillo@44 50 mahoney
meillo@44 51 oral history
meillo@1 52 .]
meillo@1 53 .IP \(bu
meillo@1 54 Mike Gancarz' book ``The UNIX Philosophy''.
meillo@1 55 .[
meillo@44 56 gancarz
meillo@44 57 unix philosophy
meillo@1 58 .]
meillo@1 59 .IP \(bu
meillo@1 60 Eric S. Raymond's book ``The Art of UNIX Programming''.
meillo@1 61 .[
meillo@44 62 raymond
meillo@44 63 art of unix programming
meillo@1 64 .]
meillo@0 65 .LP
meillo@1 66 These different views on the Unix Philosophy have much in common.
meillo@40 67 Especially, the main concepts are similar in all of them.
meillo@40 68 McIlroy's definition can surely be called the core of the Unix Philosophy,
meillo@40 69 but the fundamental idea behind it all, is ``small is beautiful''.
meillo@40 70
meillo@40 71 .PP
meillo@45 72 The Unix Philosophy explains how to design good software for Unix.
meillo@45 73 Many concepts described here, base on facilities of Unix.
meillo@40 74 Other operating systems may not offer such facilities,
meillo@41 75 hence it may not be possible to design software in the way of the
meillo@41 76 Unix Philosophy for them.
meillo@40 77 .PP
meillo@41 78 The Unix Philosophy has an idea of how the process of software development
meillo@41 79 should look like, but large parts of the philosophy are quite independent
meillo@45 80 from a concrete development process.
meillo@41 81 However, one will soon recognize that some development processes work well
meillo@41 82 with the ideas of the Unix Philosophy and support them, while others are
meillo@41 83 at cross-purposes.
meillo@45 84 Kent Beck's books about Extreme Programming are valuable supplemental
meillo@45 85 resources on this topic.
meillo@1 86 .PP
meillo@41 87 The question of how to actually write code and how the code should looks
meillo@45 88 like in detail, are out of focus here.
meillo@41 89 ``The Practice of Programming'' by Kernighan and Pike,
meillo@41 90 .[
meillo@44 91 kernighan pike
meillo@44 92 practice of programming
meillo@41 93 .]
meillo@41 94 is a good book that covers this topic.
meillo@41 95 Its point of view matches to the one of this paper.
meillo@0 96
meillo@48 97 .H 1 "Importance of software design in general
meillo@0 98 .LP
meillo@40 99 Software design is the planning of how the internal structure
meillo@40 100 and external interfaces of a software should look like.
meillo@39 101 It has nothing to do with visual appearance.
meillo@39 102 If we take a program as a car, then its color is of no matter.
meillo@39 103 Its design would be the car's size, its shape, the locations of doors,
meillo@45 104 the passenger/space ratio, the available controls and instruments,
meillo@45 105 and so forth.
meillo@39 106 .PP
meillo@39 107 Why should software get designed at all?
meillo@6 108 It is general knowledge, that even a bad plan is better than no plan.
meillo@39 109 Not designing software means programming without plan.
meillo@39 110 This will pretty sure lead to horrible results.
meillo@45 111 Software that is horrible to use and horrible to maintain.
meillo@39 112 These two aspects are the visible ones.
meillo@45 113 Often invisible though, are the wasted possible gains.
meillo@39 114 Good software design can make these gains available.
meillo@2 115 .PP
meillo@39 116 A software's design deals with quality properties.
meillo@39 117 Good design leads to good quality, and quality is important.
meillo@39 118 Any car may be able to drive from A to B,
meillo@39 119 but it depends on the car's properties whether it is a good choice
meillo@39 120 for passenger transport or not.
meillo@39 121 It depends on its properties if it is a good choice
meillo@39 122 for a rough mountain area.
meillo@39 123 And it depends on its properties if the ride will be fun.
meillo@39 124
meillo@2 125 .PP
meillo@39 126 Requirements for a software are twofold:
meillo@39 127 functional and non-functional.
meillo@39 128 .IP \(bu
meillo@39 129 Functional requirements define directly the software's functions.
meillo@39 130 They are the reason why software gets written.
meillo@39 131 Someone has a problem and needs a tool to solve it.
meillo@39 132 Being able to solve the problem is the main functional goal.
meillo@39 133 It is the driving force behind all programming effort.
meillo@39 134 Functional requirements are easier to define and to verify.
meillo@39 135 .IP \(bu
meillo@45 136 Non-functional requirements are called \fIquality\fP requirements, too.
meillo@39 137 The quality of a software are the properties that are not directly related to
meillo@39 138 the software's basic functions.
meillo@45 139 Tools of bad quality often do solve the problems they were written for,
meillo@39 140 but introduce problems and difficulties for usage and development, later on.
meillo@39 141 Quality aspects are often overlooked at first sight,
meillo@45 142 and are often difficult to define clearly and to verify.
meillo@2 143 .PP
meillo@45 144 Quality is hardly interesting when the software gets built initially,
meillo@45 145 but it has a high impact on usability and maintenance of the software, later.
meillo@45 146 A short-sighted might see in developing a software, mainly building something up.
meillo@39 147 But experience shows, that building the software the first time is
meillo@39 148 only a small amount of the overall work.
meillo@45 149 Bug fixing, extending, rebuilding of parts \(en maintenance work \(en
meillo@6 150 does soon take over the major part of the time spent on a software.
meillo@45 151 And of course, the time spent actually using the software.
meillo@6 152 These processes are highly influenced by the software's quality.
meillo@39 153 Thus, quality must not be neglected.
meillo@45 154 However, the problem with quality is that you hardly ``stumble over''
meillo@39 155 bad quality during the first build,
meillo@45 156 although this is the time when you should care about good quality most.
meillo@6 157 .PP
meillo@45 158 Software design has little to do with the basic function of a software \(en
meillo@39 159 this requirement will get satisfied anyway.
meillo@39 160 Software design is more about quality aspects of the software.
meillo@39 161 Good design leads to good quality, bad design to bad quality.
meillo@6 162 The primary functions of the software will be affected modestly by bad quality,
meillo@39 163 but good quality can provide a lot of additional gain,
meillo@6 164 even at places where one never expected it.
meillo@6 165 .PP
meillo@45 166 The ISO/IEC\|9126-1 standard, part\|1,
meillo@6 167 .[
meillo@44 168 iso product quality
meillo@6 169 .]
meillo@6 170 defines the quality model as consisting out of:
meillo@6 171 .IP \(bu
meillo@6 172 .I Functionality
meillo@6 173 (suitability, accuracy, inter\%operability, security)
meillo@6 174 .IP \(bu
meillo@6 175 .I Reliability
meillo@6 176 (maturity, fault tolerance, recoverability)
meillo@6 177 .IP \(bu
meillo@6 178 .I Usability
meillo@6 179 (understandability, learnability, operability, attractiveness)
meillo@6 180 .IP \(bu
meillo@6 181 .I Efficiency
meillo@9 182 (time behavior, resource utilization)
meillo@6 183 .IP \(bu
meillo@6 184 .I Maintainability
meillo@23 185 (analyzability, changeability, stability, testability)
meillo@6 186 .IP \(bu
meillo@6 187 .I Portability
meillo@6 188 (adaptability, installability, co-existence, replaceability)
meillo@6 189 .LP
meillo@39 190 Good design can improve these properties of a software,
meillo@45 191 bad designed software likely suffers in these points.
meillo@7 192 .PP
meillo@7 193 One further goal of software design is consistency.
meillo@7 194 Consistency eases understanding, working on, and using things.
meillo@39 195 Consistent internal structure and consistent interfaces to the outside
meillo@39 196 can be provided by good design.
meillo@7 197 .PP
meillo@39 198 Software should be well designed because good design avoids many
meillo@45 199 problems during a software's lifetime.
meillo@39 200 And software should be well designed because good design can offer
meillo@39 201 much additional gain.
meillo@39 202 Indeed, much effort should be spent into good design to make software more valuable.
meillo@39 203 The Unix Philosophy shows a way of how to design software well.
meillo@7 204 It offers guidelines to achieve good quality and high gain for the effort spent.
meillo@0 205
meillo@0 206
meillo@48 207 .H 1 "The Unix Philosophy
meillo@4 208 .LP
meillo@4 209 The origins of the Unix Philosophy were already introduced.
meillo@8 210 This chapter explains the philosophy, oriented on Gancarz,
meillo@8 211 and shows concrete examples of its application.
meillo@5 212
meillo@48 213 .H 2 Pipes
meillo@4 214 .LP
meillo@4 215 Following are some examples to demonstrate how applied Unix Philosophy feels like.
meillo@4 216 Knowledge of using the Unix shell is assumed.
meillo@4 217 .PP
meillo@4 218 Counting the number of files in the current directory:
meillo@41 219 .DS
meillo@4 220 ls | wc -l
meillo@4 221 .DE
meillo@4 222 The
meillo@4 223 .CW ls
meillo@4 224 command lists all files in the current directory, one per line,
meillo@4 225 and
meillo@4 226 .CW "wc -l
meillo@8 227 counts the number of lines.
meillo@4 228 .PP
meillo@8 229 Counting the number of files that do not contain ``foo'' in their name:
meillo@41 230 .DS
meillo@4 231 ls | grep -v foo | wc -l
meillo@4 232 .DE
meillo@4 233 Here, the list of files is filtered by
meillo@4 234 .CW grep
meillo@45 235 to remove all lines that contain ``foo''.
meillo@45 236 The rest equals the previous example.
meillo@4 237 .PP
meillo@4 238 Finding the five largest entries in the current directory.
meillo@41 239 .DS
meillo@4 240 du -s * | sort -nr | sed 5q
meillo@4 241 .DE
meillo@4 242 .CW "du -s *
meillo@45 243 returns the recursively summed sizes of all files in the current directory
meillo@8 244 \(en no matter if they are regular files or directories.
meillo@4 245 .CW "sort -nr
meillo@45 246 sorts the list numerically in reverse order (descending).
meillo@4 247 Finally,
meillo@4 248 .CW "sed 5q
meillo@4 249 quits after it has printed the fifth line.
meillo@4 250 .PP
meillo@4 251 The presented command lines are examples of what Unix people would use
meillo@4 252 to get the desired output.
meillo@4 253 There are also other ways to get the same output.
meillo@4 254 It's a user's decision which way to go.
meillo@14 255 .PP
meillo@8 256 The examples show that many tasks on a Unix system
meillo@4 257 are accomplished by combining several small programs.
meillo@4 258 The connection between the single programs is denoted by the pipe operator `|'.
meillo@4 259 .PP
meillo@4 260 Pipes, and their extensive and easy use, are one of the great
meillo@4 261 achievements of the Unix system.
meillo@4 262 Pipes between programs have been possible in earlier operating systems,
meillo@4 263 but it has never been a so central part of the concept.
meillo@45 264 When, in the early seventies, Doug McIlroy introduced pipes into the
meillo@4 265 Unix system,
meillo@4 266 ``it was this concept and notation for linking several programs together
meillo@4 267 that transformed Unix from a basic file-sharing system to an entirely new way of computing.''
meillo@4 268 .[
meillo@44 269 aughenbaugh
meillo@44 270 unix oral history
meillo@45 271 .]
meillo@4 272 .PP
meillo@4 273 Being able to specify pipelines in an easy way is,
meillo@4 274 however, not enough by itself.
meillo@5 275 It is only one half.
meillo@4 276 The other is the design of the programs that are used in the pipeline.
meillo@45 277 They need interfaces that allow them to be used in such a way.
meillo@5 278
meillo@48 279 .H 2 "Interface design
meillo@5 280 .LP
meillo@11 281 Unix is, first of all, simple \(en Everything is a file.
meillo@5 282 Files are sequences of bytes, without any special structure.
meillo@45 283 Programs should be filters, which read a stream of bytes from standard input (stdin)
meillo@45 284 and write a stream of bytes to standard output (stdout).
meillo@8 285 If the files \fIare\fP sequences of bytes,
meillo@8 286 and the programs \fIare\fP filters on byte streams,
meillo@45 287 then there is exactly one data interface.
meillo@45 288 Hence it is possible to combine programs in any desired way.
meillo@5 289 .PP
meillo@45 290 Even a handful of small programs yields a large set of combinations,
meillo@5 291 and thus a large set of different functions.
meillo@5 292 This is leverage!
meillo@5 293 If the programs are orthogonal to each other \(en the best case \(en
meillo@5 294 then the set of different functions is greatest.
meillo@5 295 .PP
meillo@45 296 Programs can also have a separate control interface,
meillo@11 297 besides their data interface.
meillo@11 298 The control interface is often called ``user interface'',
meillo@11 299 because it is usually designed to be used by humans.
meillo@11 300 The Unix Philosophy discourages to assume the user to be human.
meillo@11 301 Interactive use of software is slow use of software,
meillo@11 302 because the program waits for user input most of the time.
meillo@45 303 Interactive software requires the user to be in front of the computer.
meillo@11 304 Interactive software occupy the user's attention while they are running.
meillo@11 305 .PP
meillo@45 306 Now to come back to the idea of combining several small programs,
meillo@11 307 to have a more specific function.
meillo@11 308 If these single tools would all be interactive,
meillo@11 309 how would the user control them?
meillo@45 310 It is not only a problem to control several programs at once,
meillo@45 311 if they run at the same time,
meillo@11 312 it also very inefficient to have to control each of the single programs
meillo@45 313 that are intended to act as one large program.
meillo@11 314 Hence, the Unix Philosophy discourages programs to demand interactive use.
meillo@11 315 The behavior of programs should be defined at invocation.
meillo@45 316 This is done by specifying arguments to the program call
meillo@45 317 (command line switches).
meillo@11 318 Gancarz discusses this topic as ``avoid captive user interfaces''.
meillo@46 319 .[ [
meillo@44 320 gancarz unix philosophy
meillo@46 321 .], page 88 ff.]
meillo@11 322 .PP
meillo@11 323 Non-interactive use is, during development, also an advantage for testing.
meillo@11 324 Testing of interactive programs is much more complicated,
meillo@11 325 than testing of non-interactive programs.
meillo@5 326
meillo@48 327 .H 2 "The toolchest approach
meillo@5 328 .LP
meillo@5 329 A toolchest is a set of tools.
meillo@5 330 Instead of having one big tool for all tasks, one has many small tools,
meillo@5 331 each for one task.
meillo@5 332 Difficult tasks are solved by combining several of the small, simple tools.
meillo@5 333 .PP
meillo@11 334 The Unix toolchest \fIis\fP a set of small, (mostly) non-interactive programs
meillo@11 335 that are filters on byte streams.
meillo@11 336 They are, to a large extend, unrelated in their function.
meillo@11 337 Hence, the Unix toolchest provides a large set of functions
meillo@11 338 that can be accessed by combining the programs in the desired way.
meillo@11 339 .PP
meillo@11 340 There are also advantages for developing small toolchest programs.
meillo@5 341 It is easier and less error-prone to write small programs.
meillo@5 342 It is also easier and less error-prone to write a large set of small programs,
meillo@5 343 than to write one large program with all the functionality included.
meillo@5 344 If the small programs are combinable, then they offer even a larger set
meillo@5 345 of functions than the single large program.
meillo@45 346 Hence, one gets two advantages out of writing small, combinable programs:
meillo@45 347 They are easier to write and they offer a greater set of functions through
meillo@45 348 combination.
meillo@5 349 .PP
meillo@45 350 But there are also two main drawbacks of the toolchest approach.
meillo@45 351 First, one simple, standardized interface has to be sufficient.
meillo@5 352 If one feels the need for more ``logic'' than a stream of bytes,
meillo@8 353 then a different approach might be of need.
meillo@13 354 But it is also possible, that he just can not imagine a design where
meillo@8 355 a stream of bytes is sufficient.
meillo@8 356 By becoming more familiar with the ``Unix style of thinking'',
meillo@8 357 developers will more often and easier find simple designs where
meillo@8 358 a stream of bytes is a sufficient interface.
meillo@8 359 .PP
meillo@8 360 The second drawback of a toolchest affects the users.
meillo@45 361 A toolchest is often more difficult to use.
meillo@9 362 It is necessary to become familiar with each of the tools,
meillo@5 363 to be able to use the right one in a given situation.
meillo@45 364 Additionally, one needs to combine the tools in a senseful way himself.
meillo@45 365 This is like a sharp knife \(en it is a powerful tool in the hand of a
meillo@45 366 master, but of no good value in the hand of an unskilled.
meillo@45 367 However, learning single, small tools of a toolchest is easier than
meillo@45 368 learning a complex tool.
meillo@45 369 And the user will already have a basic understanding of a yet unknown tool,
meillo@45 370 if the tools of a toolchest have a common, consistent style.
meillo@45 371 He will be able to transfer knowledge over from one tool to another.
meillo@5 372 .PP
meillo@45 373 Moreover, the second drawback can be removed to a large extend
meillo@45 374 by adding wrappers around the basic tools.
meillo@45 375 Novice users do not need to learn several tools, if a professional wraps
meillo@45 376 complete command lines into a higher-level script.
meillo@5 377 Note that the wrapper script still calls the small tools;
meillo@45 378 it is just like a skin around them.
meillo@45 379 No complexity is added this way.
meillo@45 380 But new programs can get created out of existing one with very low effort.
meillo@5 381 .PP
meillo@5 382 A wrapper script for finding the five largest entries in the current directory
meillo@5 383 could look like this:
meillo@41 384 .DS
meillo@5 385 #!/bin/sh
meillo@5 386 du -s * | sort -nr | sed 5q
meillo@5 387 .DE
meillo@45 388 The script itself is just a text file that calls the command line,
meillo@45 389 which a professional user would type in directly.
meillo@45 390 It is probably worth to make the program flexible on the number of
meillo@45 391 entries it prints:
meillo@41 392 .DS
meillo@8 393 #!/bin/sh
meillo@8 394 num=5
meillo@8 395 [ $# -eq 1 ] && num="$1"
meillo@8 396 du -sh * | sort -nr | sed "${num}q"
meillo@8 397 .DE
meillo@8 398 This script acts like the one before, when called without an argument.
meillo@8 399 But one can also specify a numerical argument to define the number of lines to print.
meillo@45 400 One can surely imagine even more flexible versions, however,
meillo@45 401 they will still relay on the external programs,
meillo@45 402 which do the actual work.
meillo@5 403
meillo@48 404 .H 2 "A powerful shell
meillo@8 405 .LP
meillo@45 406 The Unix shell provides the possibility to combine small programs into large ones.
meillo@45 407 But a powerful shell is a great feature in other ways, too.
meillo@45 408 For instance by being scriptable.
meillo@45 409 Control statements are build into the shell.
meillo@45 410 The functions, however, are the normal programs of the system.
meillo@45 411 Thus, as the programs are already known,
meillo@45 412 learning to program in the shell becomes easy.
meillo@8 413 Using normal programs as functions in the shell programming language
meillo@10 414 is only possible because they are small and combinable tools in a toolchest style.
meillo@8 415 .PP
meillo@45 416 The Unix shell encourages to write small scripts,
meillo@45 417 by combining existing programs, because it is so easy to do.
meillo@8 418 This is a great step towards automation.
meillo@8 419 It is wonderful if the effort to automate a task equals the effort
meillo@45 420 to do the task a second time by hand.
meillo@45 421 If this holds,
meillo@45 422 then the user will be happy to automate everything he does more than once.
meillo@8 423 .PP
meillo@8 424 Small programs that do one job well, standardized interfaces between them,
meillo@8 425 a mechanism to combine parts to larger parts, and an easy way to automate tasks,
meillo@8 426 this will inevitably produce software leverage.
meillo@8 427 Getting multiple times the benefit of an investment is a great offer.
meillo@10 428 .PP
meillo@10 429 The shell also encourages rapid prototyping.
meillo@10 430 Many well known programs started as quickly hacked shell scripts,
meillo@10 431 and turned into ``real'' programs, written in C, later.
meillo@45 432 Building a prototype first, is a way to avoid the biggest problems
meillo@10 433 in application development.
meillo@45 434 Fred Brooks explains in ``No Silver Bullet'':
meillo@10 435 .[
meillo@44 436 brooks
meillo@44 437 no silver bullet
meillo@10 438 .]
meillo@10 439 .QP
meillo@10 440 The hardest single part of building a software system is deciding precisely what to build.
meillo@10 441 No other part of the conceptual work is so difficult as establishing the detailed
meillo@10 442 technical requirements, [...].
meillo@10 443 No other part of the work so cripples the resulting system if done wrong.
meillo@10 444 No other part is more difficult to rectify later.
meillo@10 445 .PP
meillo@45 446 Writing a prototype is a great method for becoming familiar with the requirements
meillo@45 447 and to run into real problems early.
meillo@47 448 .[ [
meillo@47 449 gancarz
meillo@47 450 unix philosophy
meillo@47 451 .], page 28 f.]
meillo@45 452 .PP
meillo@45 453 Prototyping is often seen as a first step in building a software.
meillo@10 454 This is, of course, good.
meillo@10 455 However, the Unix Philosophy has an \fIadditional\fP perspective on prototyping:
meillo@10 456 After having built the prototype, one might notice, that the prototype is already
meillo@10 457 \fIgood enough\fP.
meillo@45 458 Hence, no reimplementation, in a more sophisticated programming language,
meillo@45 459 might be of need, at least for the moment.
meillo@23 460 Maybe later, it might be necessary to rewrite the software, but not now.
meillo@45 461 By delaying further work, one keeps the flexibility to react on
meillo@10 462 changing requirements.
meillo@10 463 Software parts that are not written will not miss the requirements.
meillo@10 464
meillo@48 465 .H 2 "Worse is better
meillo@10 466 .LP
meillo@45 467 The Unix Philosophy aims for the 90% solution;
meillo@10 468 others call it the ``Worse is better'' approach.
meillo@47 469 Experience from real life projects shows:
meillo@10 470 .PP
meillo@45 471 (1) It is almost never possible to define the
meillo@10 472 requirements completely and correctly the first time.
meillo@45 473 Hence one should not try to; one will fail anyway.
meillo@45 474 .PP
meillo@45 475 (2) Requirements change during time.
meillo@10 476 Hence it is best to delay requirement-based design decisions as long as possible.
meillo@45 477 The software should be small and flexible as long as possible
meillo@10 478 to react on changing requirements.
meillo@10 479 Shell scripts, for example, are more easily adjusted as C programs.
meillo@45 480 .PP
meillo@45 481 (3) Maintenance work is hard work.
meillo@45 482 Hence, one should keep the amount of code as small as possible;
meillo@10 483 it should just fulfill the \fIcurrent\fP requirements.
meillo@45 484 Software parts that will be written in future,
meillo@45 485 do not need maintenance till then.
meillo@10 486 .PP
meillo@47 487 See Brooks' ``The Mythical Man-Month'' for reference.
meillo@47 488 .[ [
meillo@47 489 brooks
meillo@47 490 mythical man-month
meillo@47 491 .], page 115 ff.]
meillo@47 492 .PP
meillo@10 493 Starting with a prototype in a scripting language has several advantages:
meillo@10 494 .IP \(bu
meillo@10 495 As the initial effort is low, one will likely start right away.
meillo@10 496 .IP \(bu
meillo@10 497 As working parts are available soon, the real requirements can get identified soon.
meillo@10 498 .IP \(bu
meillo@45 499 When a software is usable and valuable, it gets used, and thus tested.
meillo@10 500 Hence problems will be found at early stages of the development.
meillo@10 501 .IP \(bu
meillo@10 502 The prototype might be enough for the moment,
meillo@45 503 thus further work on the software can get delayed to a time
meillo@10 504 when one knows better about the requirements and problems,
meillo@10 505 than now.
meillo@10 506 .IP \(bu
meillo@45 507 Implementing now only the parts that are actually needed at the moment,
meillo@45 508 introduces fewer programming and maintenance work.
meillo@10 509 .IP \(bu
meillo@10 510 If the global situation changes so that the software is not needed anymore,
meillo@10 511 then less effort was spent into the project, than it would have be
meillo@10 512 when a different approach had been used.
meillo@10 513
meillo@48 514 .H 2 "Upgrowth and survival of software
meillo@11 515 .LP
meillo@12 516 So far it was talked about \fIwriting\fP or \fIbuilding\fP software.
meillo@13 517 Although these are just verbs, they do imply a specific view on the work process
meillo@13 518 they describe.
meillo@12 519 The better verb, however, is to \fIgrow\fP.
meillo@12 520 Creating software in the sense of the Unix Philosophy is an incremental process.
meillo@12 521 It starts with a first prototype, which evolves as requirements change.
meillo@12 522 A quickly hacked shell script might become a large, sophisticated,
meillo@13 523 compiled program this way.
meillo@13 524 Its lifetime begins with the initial prototype and ends when the software is not used anymore.
meillo@45 525 While being alive it will get extended, rearranged, rebuilt.
meillo@12 526 Growing software matches the view that ``software is never finished. It is only released.''
meillo@46 527 .[ [
meillo@44 528 gancarz
meillo@44 529 unix philosophy
meillo@46 530 .], page 26]
meillo@12 531 .PP
meillo@13 532 Software can be seen as being controlled by evolutionary processes.
meillo@13 533 Successful software is software that is used by many for a long time.
meillo@12 534 This implies that the software is needed, useful, and better than alternatives.
meillo@12 535 Darwin talks about: ``The survival of the fittest.''
meillo@12 536 .[
meillo@44 537 darwin
meillo@44 538 origin of species
meillo@12 539 .]
meillo@12 540 Transferred to software: The most successful software, is the fittest,
meillo@12 541 is the one that survives.
meillo@13 542 (This may be at the level of one creature, or at the level of one species.)
meillo@13 543 The fitness of software is affected mainly by four properties:
meillo@15 544 portability of code, portability of data, range of usability, and reusability of parts.
meillo@13 545 .PP
meillo@15 546 (1)
meillo@15 547 .I "Portability of code
meillo@15 548 means, using high-level programming languages,
meillo@13 549 sticking to the standard,
meillo@47 550 .[ [
meillo@47 551 kernighan pike
meillo@47 552 practice of programming
meillo@47 553 .], chapter\|8]
meillo@13 554 and avoiding optimizations that introduce dependencies on specific hardware.
meillo@13 555 Hardware has a much lower lifetime than software.
meillo@13 556 By chaining software to a specific hardware,
meillo@13 557 the software's lifetime gets shortened to that of this hardware.
meillo@13 558 In contrast, software should be easy to port \(en
meillo@23 559 adaptation is the key to success.
meillo@13 560 .PP
meillo@15 561 (2)
meillo@15 562 .I "Portability of data
meillo@15 563 is best achieved by avoiding binary representations
meillo@13 564 to store data, because binary representations differ from machine to machine.
meillo@23 565 Textual representation is favored.
meillo@45 566 Historically, \s-1ASCII\s0 was the charset of choice.
meillo@45 567 For the future, \s-1UTF\s0-8 might be the better choice.
meillo@13 568 Important is that it is a plain text representation in a
meillo@13 569 very common charset encoding.
meillo@13 570 Apart from being able to transfer data between machines,
meillo@45 571 readable data has the great advantage, that humans are able to directly
meillo@45 572 read and edit it with text editors and other tools from the Unix toolchest.
meillo@47 573 .[ [
meillo@47 574 gancarz
meillo@47 575 unix philosophy
meillo@47 576 .], page 56 ff.]
meillo@13 577 .PP
meillo@15 578 (3)
meillo@15 579 A large
meillo@15 580 .I "range of usability
meillo@23 581 ensures good adaptation, and thus good survival.
meillo@13 582 It is a special distinction if a software becomes used in fields of action,
meillo@13 583 the original authors did never imagine.
meillo@13 584 Software that solves problems in a general way will likely be used
meillo@45 585 for many kinds of similar problems.
meillo@45 586 Being too specific limits the range of usability.
meillo@13 587 Requirements change through time, thus use cases change or even vanish.
meillo@45 588 As a good example in this point,
meillo@13 589 Allman identifies flexibility to be one major reason for sendmail's success:
meillo@13 590 .[
meillo@44 591 allman
meillo@44 592 sendmail
meillo@13 593 .]
meillo@13 594 .QP
meillo@13 595 Second, I limited myself to the routing function [...].
meillo@13 596 This was a departure from the dominant thought of the time, [...].
meillo@13 597 .QP
meillo@45 598 Third, the sendmail configuration file was flexible enough to adapt
meillo@13 599 to a rapidly changing world [...].
meillo@12 600 .LP
meillo@45 601 Successful software adapts itself to the changing world.
meillo@13 602 .PP
meillo@15 603 (4)
meillo@15 604 .I "Reuse of parts
meillo@15 605 is even one step further.
meillo@13 606 A software may completely lose its field of action,
meillo@13 607 but parts of which the software is build may be general and independent enough
meillo@13 608 to survive this death.
meillo@13 609 If software is build by combining small independent programs,
meillo@45 610 then these parts are readily available for reuse.
meillo@13 611 Who cares if the large program is a failure,
meillo@13 612 but parts of it become successful instead?
meillo@10 613
meillo@48 614 .H 2 "Summary
meillo@0 615 .LP
meillo@45 616 This chapter explained central ideas of the Unix Philosophy.
meillo@45 617 For each of the ideas, the advantages they introduce were explained.
meillo@45 618 The Unix Philosophy are guidelines that help to write more valuable software.
meillo@14 619 From the view point of a software developer or software designer,
meillo@14 620 the Unix Philosophy provides answers to many software design problem.
meillo@14 621 .PP
meillo@14 622 The various ideas of the Unix Philosophy are very interweaved
meillo@14 623 and can hardly be applied independently.
meillo@14 624 However, the probably most important messages are:
meillo@45 625 .I "``Keep it simple!''" ,
meillo@14 626 .I "``Do one thing well!''" ,
meillo@14 627 and
meillo@14 628 .I "``Use software leverage!''
meillo@0 629
meillo@8 630
meillo@8 631
meillo@48 632 .H 1 "Case study: \s-1MH\s0
meillo@18 633 .LP
meillo@30 634 The previous chapter introduced and explained the Unix Philosophy
meillo@18 635 from a general point of view.
meillo@30 636 The driving force were the guidelines; references to
meillo@18 637 existing software were given only sparsely.
meillo@18 638 In this and the next chapter, concrete software will be
meillo@18 639 the driving force in the discussion.
meillo@18 640 .PP
meillo@23 641 This first case study is about the mail user agents (\s-1MUA\s0)
meillo@23 642 \s-1MH\s0 (``mail handler'') and its descendent \fInmh\fP
meillo@23 643 (``new mail handler'').
meillo@47 644 .[
meillo@47 645 nmh website
meillo@47 646 .]
meillo@23 647 \s-1MUA\s0s provide functions to read, compose, and organize mail,
meillo@45 648 but (ideally) not to transfer it.
meillo@45 649 In this document, the name \s-1MH\s0 will be used to include nmh.
meillo@19 650 A distinction will only be made if differences between
meillo@45 651 \s-1MH\s0 and nmh are described.
meillo@18 652
meillo@0 653
meillo@48 654 .H 2 "Historical background
meillo@0 655 .LP
meillo@19 656 Electronic mail was available in Unix very early.
meillo@30 657 The first \s-1MUA\s0 on Unix was \f(CWmail\fP,
meillo@30 658 which was already present in the First Edition.
meillo@46 659 .[ [
meillo@44 660 salus
meillo@44 661 quarter century of unix
meillo@46 662 .], page 41 f.]
meillo@45 663 It was a small program that either printed the user's mailbox file
meillo@45 664 or appended text to someone elses mailbox file,
meillo@19 665 depending on the command line arguments.
meillo@19 666 .[
meillo@44 667 manual mail(1)
meillo@19 668 .]
meillo@19 669 It was a program that did one job well.
meillo@23 670 This job was emailing, which was very simple then.
meillo@19 671 .PP
meillo@23 672 Later, emailing became more powerful, and thus more complex.
meillo@19 673 The simple \f(CWmail\fP, which knew nothing of subjects,
meillo@19 674 independent handling of single messages,
meillo@45 675 and long-time email storage, was not powerful enough anymore.
meillo@45 676 In 1978 at Berkeley, Kurt Shoens wrote \fIMail\fP (with capital `M')
meillo@45 677 to provide additional functions for emailing.
meillo@19 678 Mail was still one program, but now it was large and did
meillo@19 679 several jobs.
meillo@23 680 Its user interface is modeled after the one of \fIed\fP.
meillo@19 681 It is designed for humans, but is still scriptable.
meillo@23 682 \fImailx\fP is the adaptation of Berkeley Mail into System V.
meillo@19 683 .[
meillo@44 684 ritter
meillo@44 685 mailx history
meillo@19 686 .]
meillo@30 687 Elm, pine, mutt, and a whole bunch of graphical \s-1MUA\s0s
meillo@19 688 followed Mail's direction.
meillo@19 689 They are large, monolithic programs which include all emailing functions.
meillo@19 690 .PP
meillo@23 691 A different way was taken by the people of \s-1RAND\s0 Corporation.
meillo@38 692 In the beginning, they also had used a monolithic mail system,
meillo@30 693 called \s-1MS\s0 (for ``mail system'').
meillo@19 694 But in 1977, Stockton Gaines and Norman Shapiro
meillo@19 695 came up with a proposal of a new email system concept \(en
meillo@45 696 one that honored the Unix Philosophy.
meillo@19 697 The concept was implemented by Bruce Borden in 1978 and 1979.
meillo@19 698 This was the birth of \s-1MH\s0 \(en the ``mail handler''.
meillo@18 699 .PP
meillo@18 700 Since then, \s-1RAND\s0, the University of California at Irvine and
meillo@19 701 at Berkeley, and several others have contributed to the software.
meillo@18 702 However, it's core concepts remained the same.
meillo@23 703 In the late 90s, when development of \s-1MH\s0 slowed down,
meillo@19 704 Richard Coleman started with \fInmh\fP, the new mail handler.
meillo@45 705 His goal was to improve \s-1MH\s0 especially in regard of
meillo@23 706 the requirements of modern emailing.
meillo@19 707 Today, nmh is developed by various people on the Internet.
meillo@18 708 .[
meillo@44 709 ware
meillo@44 710 rand history
meillo@18 711 .]
meillo@18 712 .[
meillo@44 713 peek
meillo@44 714 mh
meillo@18 715 .]
meillo@0 716
meillo@48 717 .H 2 "Contrasts to monolithic mail systems
meillo@0 718 .LP
meillo@19 719 All \s-1MUA\s0s are monolithic, except \s-1MH\s0.
meillo@38 720 Although there might actually exist further, very little known,
meillo@30 721 toolchest \s-1MUA\s0s, this statement reflects the situation pretty well.
meillo@19 722 .PP
meillo@30 723 Monolithic \s-1MUA\s0s gather all their functions in one program.
meillo@30 724 In contrast, \s-1MH\s0 is a toolchest of many small tools \(en one for each job.
meillo@23 725 Following is a list of important programs of \s-1MH\s0's toolchest
meillo@30 726 and their function.
meillo@30 727 It gives a feeling of how the toolchest looks like.
meillo@19 728 .IP \(bu
meillo@19 729 .CW inc :
meillo@30 730 incorporate new mail (this is how mail enters the system)
meillo@19 731 .IP \(bu
meillo@19 732 .CW scan :
meillo@19 733 list messages in folder
meillo@19 734 .IP \(bu
meillo@19 735 .CW show :
meillo@19 736 show message
meillo@19 737 .IP \(bu
meillo@19 738 .CW next\fR/\fPprev :
meillo@19 739 show next/previous message
meillo@19 740 .IP \(bu
meillo@19 741 .CW folder :
meillo@19 742 change current folder
meillo@19 743 .IP \(bu
meillo@19 744 .CW refile :
meillo@45 745 refile message into different folder
meillo@19 746 .IP \(bu
meillo@19 747 .CW rmm :
meillo@19 748 remove message
meillo@19 749 .IP \(bu
meillo@19 750 .CW comp :
meillo@45 751 compose new message
meillo@19 752 .IP \(bu
meillo@19 753 .CW repl :
meillo@45 754 reply to message
meillo@19 755 .IP \(bu
meillo@19 756 .CW forw :
meillo@45 757 forward message
meillo@19 758 .IP \(bu
meillo@19 759 .CW send :
meillo@45 760 send prepared message (this is how mail leaves the system)
meillo@0 761 .LP
meillo@19 762 \s-1MH\s0 has no special user interface like monolithic \s-1MUA\s0s have.
meillo@19 763 The user does not leave the shell to run \s-1MH\s0,
meillo@45 764 instead he uses the various \s-1MH\s0 programs within the shell.
meillo@23 765 Using a monolithic program with a captive user interface
meillo@23 766 means ``entering'' the program, using it, and ``exiting'' the program.
meillo@23 767 Using toolchests like \s-1MH\s0 means running programs,
meillo@45 768 alone or in combination with others, also from other toolchests,
meillo@23 769 without leaving the shell.
meillo@30 770
meillo@48 771 .H 2 "Data storage
meillo@30 772 .LP
meillo@34 773 \s-1MH\s0's mail storage is a directory tree under the user's
meillo@34 774 \s-1MH\s0 directory (usually \f(CW$HOME/Mail\fP),
meillo@34 775 where mail folders are directories and mail messages are text files
meillo@34 776 within them.
meillo@34 777 Each mail folder contains a file \f(CW.mh_sequences\fP which lists
meillo@45 778 the public message sequences of that folder,
meillo@45 779 for instance the \fIunseen\fP sequence for new messages.
meillo@34 780 Mail messages are text files located in a mail folder.
meillo@34 781 The files contain the messages as they were received.
meillo@45 782 They are named by ascending numbers in each folder.
meillo@19 783 .PP
meillo@30 784 This mailbox format is called ``\s-1MH\s0'' after the \s-1MUA\s0.
meillo@30 785 Alternatives are \fImbox\fP and \fImaildir\fP.
meillo@30 786 In the mbox format all messages are stored within one file.
meillo@30 787 This was a good solution in the early days, when messages
meillo@30 788 were only a few lines of text and were deleted soon.
meillo@30 789 Today, when single messages often include several megabytes
meillo@30 790 of attachments, it is a bad solution.
meillo@30 791 Another disadvantage of the mbox format is that it is
meillo@30 792 more difficult to write tools that work on mail messages,
meillo@30 793 because it is always necessary to first find and extract
meillo@30 794 the relevant message in the mbox file.
meillo@45 795 With the \s-1MH\s0 mailbox format, each message is a separate file.
meillo@30 796 Also, the problem of concurrent access to one mailbox is
meillo@30 797 reduced to the problem of concurrent access to one message.
meillo@45 798 The maildir format is generally similar to the \s-1MH\s0 format,
meillo@30 799 but modified towards guaranteed reliability.
meillo@30 800 This involves some complexity, unfortunately.
meillo@34 801 .PP
meillo@34 802 Working with \s-1MH\s0's toolchest on mailboxes is much like
meillo@34 803 working with Unix' toolchest on directory trees:
meillo@34 804 \f(CWscan\fP is like \f(CWls\fP,
meillo@34 805 \f(CWshow\fP is like \f(CWcat\fP,
meillo@34 806 \f(CWfolder\fP is like \f(CWcd\fP and \f(CWpwd\fP,
meillo@34 807 \f(CWrefile\fP is like \f(CWmv\fP,
meillo@34 808 and \f(CWrmm\fP is like \f(CWrm\fP.
meillo@34 809 .PP
meillo@45 810 \s-1MH\s0 extends the context of processes in Unix by two more items,
meillo@45 811 for its tools:
meillo@34 812 .IP \(bu
meillo@34 813 The current mail folder, which is similar to the current working directory.
meillo@34 814 For mail folders, \f(CWfolder\fP provides the corresponding functionality
meillo@34 815 of \f(CWcd\fP and \f(CWpwd\fP for directories.
meillo@34 816 .IP \(bu
meillo@34 817 Sequences, which are named sets of messages in a mail folder.
meillo@34 818 The current message, relative to a mail folder, is a special sequence.
meillo@34 819 It enables commands like \f(CWnext\fP and \f(CWprev\fP.
meillo@34 820 .LP
meillo@45 821 In contrast to Unix' context, which is maintained by the kernel,
meillo@45 822 \s-1MH\s0's context must be maintained by the tools themselves.
meillo@45 823 Usually there is one context per user, which resides in his
meillo@45 824 \f(CWcontext\fP file in the \s-1MH\s0 directory,
meillo@45 825 but a user can have several contexts, too.
meillo@45 826 Public sequences are an exception, as they belong to a mail folder,
meillo@45 827 and reside in the \f(CW.mh_sequences\fP file there.
meillo@34 828 .[
meillo@44 829 man page mh-profile mh-sequence
meillo@34 830 .]
meillo@20 831
meillo@48 832 .H 2 "Discussion of the design
meillo@0 833 .LP
meillo@45 834 This section discusses \s-1MH\s0 in regard to the tenets
meillo@45 835 of the Unix Philosophy that Gancarz identified.
meillo@20 836
meillo@20 837 .PP
meillo@33 838 .B "Small is beautiful
meillo@20 839 and
meillo@33 840 .B "do one thing well
meillo@20 841 are two design goals that are directly visible in \s-1MH\s0.
meillo@45 842 Gancarz actually presents \s-1MH\s0 in his book as example under the
meillo@45 843 headline ``Making \s-1UNIX\s0 Do One Thing Well'':
meillo@46 844 .[ [
meillo@44 845 gancarz
meillo@44 846 unix philosophy
meillo@46 847 .], page 125 ff.]
meillo@20 848 .QP
meillo@20 849 [\s-1MH\s0] consists of a series of programs which
meillo@20 850 when combined give the user an enormous ability
meillo@20 851 to manipulate electronic mail messages.
meillo@20 852 A complex application, it shows that not only is it
meillo@20 853 possible to build large applications from smaller
meillo@20 854 components, but also that such designs are actually preferable.
meillo@20 855 .LP
meillo@45 856 The various programs of \s-1MH\s0 were relatively easy to write,
meillo@45 857 because each of them is small, limited to one function,
meillo@23 858 and has clear boundaries.
meillo@20 859 For the same reasons, they are also good to maintain.
meillo@20 860 Further more, the system can easily get extended.
meillo@20 861 One only needs to put a new program into the toolchest.
meillo@23 862 This was done, for instance, when \s-1MIME\s0 support was added
meillo@20 863 (e.g. \f(CWmhbuild\fP).
meillo@20 864 Also, different programs can exist to do the basically same job
meillo@20 865 in different ways (e.g. in nmh: \f(CWshow\fP and \f(CWmhshow\fP).
meillo@45 866 .PP
meillo@20 867 If someone needs a mail system with some additionally
meillo@45 868 functions that are not available anywhere yet, he best expands a
meillo@45 869 toolchest system like \s-1MH\s0.
meillo@45 870 There he can add new functionality by simply adding additional
meillo@45 871 programs to the toolchest.
meillo@45 872 There he does not risk to break existing functionality by doing so.
meillo@20 873
meillo@20 874 .PP
meillo@34 875 .B "Store data in flat text files
meillo@34 876 is followed by \s-1MH\s0.
meillo@34 877 This is not surprising, because email messages are already plain text.
meillo@34 878 \s-1MH\s0 stores the messages as it receives them,
meillo@45 879 thus any other tool that works on \s-1RFC\s0\|2822 mail messages can operate
meillo@34 880 on the messages in an \s-1MH\s0 mailbox.
meillo@45 881 All other files \s-1MH\s0 uses are plain text, too.
meillo@34 882 It is therefore possible and encouraged to use the text processing
meillo@34 883 tools of Unix' toolchest to extend \s-1MH\s0's toolchest.
meillo@20 884
meillo@20 885 .PP
meillo@33 886 .B "Avoid captive user interfaces" .
meillo@19 887 \s-1MH\s0 is perfectly suited for non-interactive use.
meillo@19 888 It offers all functions directly and without captive user interfaces.
meillo@30 889 If, nonetheless, users want a graphical user interface,
meillo@45 890 they can have it with \fIxmh\fP or \fIexmh\fP.
meillo@19 891 These are graphical frontends for the \s-1MH\s0 toolchest.
meillo@19 892 This means, all email-related work is still done by \s-1MH\s0 tools,
meillo@45 893 but the frontend calls the appropriate commands when the user
meillo@30 894 clicks on buttons.
meillo@45 895 .PP
meillo@20 896 Providing easy-to-use user interfaces in form of frontends is a good
meillo@19 897 approach, because it does not limit the power of the backend itself.
meillo@20 898 The frontend will anyway only be able to make a subset of the
meillo@23 899 backend's power and flexibility available to the user.
meillo@20 900 But if it is a separate program,
meillo@20 901 then the missing parts can still be accessed at the backend directly.
meillo@19 902 If it is integrated, then this will hardly be possible.
meillo@45 903 An additional advantage is the possibility to have different frontends
meillo@45 904 to the same backend.
meillo@19 905
meillo@19 906 .PP
meillo@33 907 .B "Choose portability over efficiency
meillo@20 908 and
meillo@33 909 .B "use shell scripts to increase leverage and portability" .
meillo@20 910 These two tenets are indirectly, but nicely, demonstrated by
meillo@30 911 Bolsky and Korn in their book about the Korn Shell.
meillo@20 912 .[
meillo@44 913 bolsky korn
meillo@44 914 korn shell
meillo@20 915 .]
meillo@45 916 Chapter\|18 of the book shows a basic implementation
meillo@20 917 of a subset of \s-1MH\s0 in ksh scripts.
meillo@45 918 Of course, this is just a demonstration, but a brilliant one.
meillo@20 919 It shows how quickly one can implement such a prototype with shell scripts,
meillo@20 920 and how readable they are.
meillo@20 921 The implementation in the scripting language may not be very fast,
meillo@20 922 but it can be fast enough though, and this is all that matters.
meillo@20 923 By having the code in an interpreted language, like the shell,
meillo@20 924 portability becomes a minor issue, if we assume the interpreter
meillo@20 925 to be widespread.
meillo@45 926 .PP
meillo@20 927 This demonstration also shows how easy it is to create single programs
meillo@20 928 of a toolchest software.
meillo@45 929 Eight tools (two of them have multiple names) and 16 functions
meillo@45 930 with supporting code are presented to the reader.
meillo@45 931 The tools comprise less than 40 lines of ksh each,
meillo@30 932 in total about 200 lines.
meillo@45 933 The functions comprise less than 80 lines of ksh each,
meillo@30 934 in total about 450 lines.
meillo@20 935 Such small software is easy to write, easy to understand,
meillo@20 936 and thus easy to maintain.
meillo@23 937 A toolchest improves the possibility to only write some parts
meillo@20 938 and though create a working result.
meillo@45 939 Expanding the toolchest, even without global changes,
meillo@45 940 will likely be possible.
meillo@20 941
meillo@20 942 .PP
meillo@33 943 .B "Use software leverage to your advantage
meillo@20 944 and the lesser tenet
meillo@33 945 .B "allow the user to tailor the environment
meillo@20 946 are ideally followed in the design of \s-1MH\s0.
meillo@21 947 Tailoring the environment is heavily encouraged by the ability to
meillo@30 948 directly define default options to programs.
meillo@30 949 It is even possible to define different default options
meillo@45 950 depending on the name under which a program is called.
meillo@45 951 Software leverage is heavily encouraged by the ease of
meillo@45 952 creating shell scripts that run a specific command line,
meillo@30 953 built of several \s-1MH\s0 programs.
meillo@21 954 There is few software that so much wants users to tailor their
meillo@21 955 environment and to leverage the use of the software, like \s-1MH\s0.
meillo@45 956 .PP
meillo@21 957 Just to make one example:
meillo@23 958 One might prefer a different listing format for the \f(CWscan\fP
meillo@21 959 program.
meillo@30 960 It is possible to take one of the distributed format files
meillo@21 961 or to write one yourself.
meillo@21 962 To use the format as default for \f(CWscan\fP, a single line,
meillo@21 963 reading
meillo@21 964 .DS
meillo@21 965 scan: -form FORMATFILE
meillo@21 966 .DE
meillo@21 967 must be added to \f(CW.mh_profile\fP.
meillo@21 968 If one wants this different format as an additional command,
meillo@23 969 instead of changing the default, he needs to create a link to
meillo@23 970 \f(CWscan\fP, for instance titled \f(CWscan2\fP.
meillo@21 971 The line in \f(CW.mh_profile\fP would then start with \f(CWscan2\fP,
meillo@45 972 as the option should only be in effect for a program that is called as
meillo@21 973 \f(CWscan2\fP.
meillo@20 974
meillo@20 975 .PP
meillo@33 976 .B "Make every program a filter
meillo@21 977 is hard to find in \s-1MH\s0.
meillo@21 978 The reason therefore is that most of \s-1MH\s0's tools provide
meillo@45 979 basic file system operations for mailboxes.
meillo@45 980 It is the same reason because of which \f(CWls\fP, \f(CWcp\fP, \f(CWmv\fP,
meillo@45 981 and \f(CWrm\fP aren't filters neither.
meillo@23 982 \s-1MH\s0 does not provide many filters itself, but it is a basis
meillo@23 983 to write filters for.
meillo@45 984 An example would be a mail text highlighter,
meillo@30 985 that means a program that makes use of a color terminal to display
meillo@30 986 header lines, quotations, and signatures in distinct colors.
meillo@45 987 The author's version of such a program is an awk script with 25 lines.
meillo@21 988
meillo@21 989 .PP
meillo@33 990 .B "Build a prototype as soon as possible
meillo@21 991 was again well followed by \s-1MH\s0.
meillo@21 992 This tenet, of course, focuses on early development, which is
meillo@21 993 long time ago for \s-1MH\s0.
meillo@21 994 But without following this guideline at the very beginning,
meillo@23 995 Bruce Borden may have not convinced the management of \s-1RAND\s0
meillo@23 996 to ever create \s-1MH\s0.
meillo@23 997 In Bruce' own words:
meillo@46 998 .[ [
meillo@44 999 ware rand history
meillo@46 1000 .], page 132]
meillo@21 1001 .QP
meillo@45 1002 [...] but [Stockton Gaines and Norm Shapiro] were not able
meillo@23 1003 to convince anyone that such a system would be fast enough to be usable.
meillo@21 1004 I proposed a very short project to prove the basic concepts,
meillo@21 1005 and my management agreed.
meillo@21 1006 Looking back, I realize that I had been very lucky with my first design.
meillo@21 1007 Without nearly enough design work,
meillo@21 1008 I built a working environment and some header files
meillo@21 1009 with key structures and wrote the first few \s-1MH\s0 commands:
meillo@21 1010 inc, show/next/prev, and comp.
meillo@21 1011 [...]
meillo@21 1012 With these three, I was able to convince people that the structure was viable.
meillo@21 1013 This took about three weeks.
meillo@0 1014
meillo@48 1015 .H 2 "Problems
meillo@0 1016 .LP
meillo@45 1017 \s-1MH\s0 is not without problems.
meillo@30 1018 There are two main problems: one is technical, the other is about human behavior.
meillo@22 1019 .PP
meillo@22 1020 \s-1MH\s0 is old and email today is very different to email in the time
meillo@22 1021 when \s-1MH\s0 was designed.
meillo@45 1022 \s-1MH\s0 adapted to the changes pretty well, but it is limited, though.
meillo@22 1023 \s-1MIME\s0 support and support for different character encodings
meillo@22 1024 is available, but only on a moderate level.
meillo@45 1025 This comes from limited development resources.
meillo@45 1026 More active developers could quickly change this.
meillo@45 1027 But \s-1MH\s0 is also limited by design, which is the larger problem.
meillo@22 1028 \s-1IMAP\s0, for example, conflicts with \s-1MH\s0's design to a large extend.
meillo@22 1029 These design conflicts are not easily solvable.
meillo@22 1030 Possibly, they require a redesign.
meillo@45 1031 \s-1IMAP\s0 may be too different to the classic mail model,
meillo@45 1032 which \s-1MH\s0 covers, so that \s-1MH\s0 may never support it well.
meillo@22 1033 .PP
meillo@45 1034 The other kind of problem are human habits.
meillo@45 1035 In this world, where almost all \s-1MUA\s0s are monolithic,
meillo@22 1036 it is very difficult to convince people to use a toolbox style \s-1MUA\s0
meillo@22 1037 like \s-1MH\s0.
meillo@45 1038 The habits are so strong, that even people who understand the concept
meillo@30 1039 and advantages of \s-1MH\s0 do not like to switch,
meillo@30 1040 simply because \s-1MH\s0 is different.
meillo@30 1041 Unfortunately, the frontends to \s-1MH\s0, which could provide familiar look'n'feel,
meillo@45 1042 are quite outdated and thus not very appealing, compared to the modern interfaces
meillo@45 1043 of many monolithic \s-1MUA\s0s.
meillo@20 1044
meillo@48 1045 NH 2 "Summary \s-1MH\s0
meillo@20 1046 .LP
meillo@45 1047 \s-1MH\s0 is an \s-1MUA\s0 that follows the Unix Philosophy in its design.
meillo@31 1048 It consists of a toolchest of small tools, each of them does one job well.
meillo@31 1049 The toolchest approach offers great flexibility to the user.
meillo@45 1050 It is possible to utilize the complete power of the Unix shell with \s-1MH\s0.
meillo@31 1051 This makes \s-1MH\s0 a very powerful mail system.
meillo@45 1052 Extending and customizing \s-1MH\s0 is easy and encouraged.
meillo@31 1053 .PP
meillo@31 1054 Apart from the user's perspective, \s-1MH\s0 is development-friendly.
meillo@31 1055 Its overall design follows clear rules.
meillo@31 1056 The single tools do only one job, thus they are easy to understand,
meillo@31 1057 easy to write, and good to maintain.
meillo@31 1058 They are all independent and do not interfere with the others.
meillo@31 1059 Automated testing of their function is a straight forward task.
meillo@31 1060 .PP
meillo@31 1061 It is sad, that \s-1MH\s0's differentness is its largest problem,
meillo@31 1062 as its differentness is also its largest advantage.
meillo@31 1063 Unfortunately, for most people their habits are stronger
meillo@31 1064 than the attraction of the clear design and the power, \s-1MH\s0 offers.
meillo@0 1065
meillo@8 1066
meillo@8 1067
meillo@48 1068 .H 1 "Case study: uzbl
meillo@32 1069 .LP
meillo@32 1070 The last chapter took a look on the \s-1MUA\s0 \s-1MH\s0,
meillo@45 1071 which is an old and established software.
meillo@45 1072 This chapter covers uzbl, a fresh new project.
meillo@45 1073 Uzbl is a web browser that adheres to the Unix Philosophy.
meillo@45 1074 Its name comes from the \fILolspeak\fP word for ``usable'';
meillo@45 1075 it is pronounced identical.
meillo@0 1076
meillo@48 1077 .H 2 "Historical background
meillo@0 1078 .LP
meillo@32 1079 Uzbl was started by Dieter Plaetinck in April 2009.
meillo@44 1080 The idea was born in a thread in the Arch Linux Forums.
meillo@32 1081 .[
meillo@44 1082 arch linux forums
meillo@44 1083 browser
meillo@32 1084 .]
meillo@32 1085 After some discussion about failures of well known web browsers,
meillo@32 1086 Plaetinck (alias Dieter@be) came up with a very sketchy proposal
meillo@32 1087 of how a better web browser could look like.
meillo@32 1088 To the question of another member, if Plaetinck would write that program,
meillo@32 1089 because it would sound fantastic, Plaetinck replied:
meillo@32 1090 ``Maybe, if I find the time ;-)''.
meillo@32 1091 .PP
meillo@32 1092 Fortunately, he found the time.
meillo@32 1093 One day later, the first prototype was out.
meillo@32 1094 One week later, uzbl had an own website.
meillo@47 1095 .[
meillo@47 1096 uzbl website
meillo@47 1097 .]
meillo@32 1098 One month after the first code showed up,
meillo@47 1099 a mailing list was installed to coordinate and discuss further development,
meillo@47 1100 and a wiki was added to store documentation and scripts that showed up on the
meillo@32 1101 mailing list and elsewhere.
meillo@32 1102 .PP
meillo@45 1103 In the, now, one year of uzbl's existence, it was heavily developed on various branches.
meillo@32 1104 Plaetinck's task became more and more to only merge the best code from the
meillo@32 1105 different branches into his main branch, and to apply patches.
meillo@47 1106 .[
meillo@47 1107 lwn uzbl
meillo@47 1108 .]
meillo@32 1109 About once a month, Plaetinck released a new version.
meillo@32 1110 In September 2009, he presented several forks of uzbl.
meillo@47 1111 .[ [
meillo@47 1112 uzbl website
meillo@47 1113 .], news archive]
meillo@38 1114 Uzbl, actually, opened the field for a whole family of web browsers with similar shape.
meillo@32 1115 .PP
meillo@32 1116 In July 2009, \fILinux Weekly News\fP published an interview with Plaetinck about uzbl.
meillo@47 1117 .[
meillo@47 1118 lwn uzbl
meillo@47 1119 .]
meillo@32 1120 In September 2009, the uzbl web browser was on \fISlashdot\fP.
meillo@47 1121 .[
meillo@47 1122 slashdot uzbl
meillo@47 1123 .]
meillo@0 1124
meillo@48 1125 .H 2 "Contrasts to other web browsers
meillo@0 1126 .LP
meillo@32 1127 Like most \s-1MUA\s0s are monolithic, but \s-1MH\s0 is a toolchest,
meillo@32 1128 most web browsers are monolithic, but uzbl is a frontend to a toolchest.
meillo@32 1129 .PP
meillo@32 1130 Today, uzbl is divided into uzbl-core and uzbl-browser.
meillo@32 1131 Uzbl-core is, how its name already indicates, the core of uzbl.
meillo@32 1132 It handles commands and events to interface other programs,
meillo@45 1133 and also displays webpages by using \fIwebkit\fP as render engine.
meillo@32 1134 Uzbl-browser combines uzbl-core with a bunch of handler scripts, a status bar,
meillo@32 1135 an event manager, yanking, pasting, page searching, zooming, and more stuff,
meillo@32 1136 to form a ``complete'' web browser.
meillo@32 1137 In the following text, the term ``uzbl'' usually stands for uzbl-browser,
meillo@32 1138 so uzbl-core is included.
meillo@32 1139 .PP
meillo@32 1140 Unlike most other web browsers, uzbl is mainly the mediator between the
meillo@45 1141 various tools that cover single jobs.
meillo@35 1142 Therefore, uzbl listens for commands on a named pipe (fifo), a Unix socket,
meillo@35 1143 and on stdin, and it writes events to a Unix socket and to stdout.
meillo@35 1144 Loading a webpage in a running uzbl instance requires only:
meillo@32 1145 .DS
meillo@32 1146 echo 'uri http://example.org' >/path/to/uzbl-fifo
meillo@32 1147 .DE
meillo@45 1148 The graphical rendering of the webpage is done by webkit,
meillo@45 1149 a web content engine.
meillo@45 1150 Uzbl-core is built around libwebkit.
meillo@32 1151 .PP
meillo@45 1152 Downloads, browsing history, bookmarks, and the like are not provided
meillo@45 1153 by the core itself, like they are in other web browsers.
meillo@35 1154 Uzbl-browser also only provides, so called, handler scripts that wrap
meillo@35 1155 external applications which provide the actual functionality.
meillo@32 1156 For instance, \fIwget\fP is used to download files and uzbl-browser
meillo@32 1157 includes a script that calls wget with appropriate options in
meillo@32 1158 a prepared environment.
meillo@32 1159 .PP
meillo@32 1160 Modern web browsers are proud to have addons, plugins, and modules, instead.
meillo@32 1161 This is their effort to achieve similar goals.
meillo@35 1162 But instead of using existing, external programs, modern web browsers
meillo@45 1163 include these functions.
meillo@0 1164
meillo@48 1165 .H 2 "Discussion of the design
meillo@0 1166 .LP
meillo@32 1167 This section discusses uzbl in regard of the Unix Philosophy,
meillo@32 1168 as identified by Gancarz.
meillo@32 1169
meillo@32 1170 .PP
meillo@35 1171 .B "Make each program do one thing well" .
meillo@35 1172 Uzbl tries to be a web browser and nothing else.
meillo@36 1173 The common definition of a web browser is, of course, highly influenced by
meillo@36 1174 existing implementations of web browsers, although they are degenerated.
meillo@35 1175 Web browsers should be programs to browse the web, and nothing more.
meillo@45 1176 This is the one thing they should do.
meillo@36 1177 .PP
meillo@45 1178 Web browsers should not, for instance, manage downloads.
meillo@35 1179 This is the job download managers exist for.
meillo@35 1180 Download managers do primary care about being good in downloading files.
meillo@35 1181 Modern web browsers provide download management only as a secondary feature.
meillo@45 1182 How could they do this job better, than programs that exist only for
meillo@35 1183 this very job?
meillo@35 1184 And how could anyone want less than the best download manager available?
meillo@32 1185 .PP
meillo@35 1186 A web browser's job is to let the user browse the web.
meillo@35 1187 This means, navigating through websites by following links.
meillo@36 1188 Rendering the \s-1HTML\s0 sources is a different job, too.
meillo@36 1189 It is covered by the webkit render engine, in uzbl's case.
meillo@35 1190 Audio and video content and files like PostScript, \s-1PDF\s0, and the like,
meillo@36 1191 are also not the job of a web browser.
meillo@36 1192 They should be handled by external applications \(en
meillo@36 1193 ones which's job is to handle such data.
meillo@35 1194 Uzbl strives to do it this way.
meillo@36 1195 .PP
meillo@36 1196 Remember Doug McIlroy:
meillo@35 1197 .I
meillo@35 1198 ``Write programs that do one thing and do it well.
meillo@35 1199 Write programs to work together.''
meillo@35 1200 .R
meillo@35 1201 .PP
meillo@35 1202 The lesser tenet
meillo@35 1203 .B "allow the user to tailor the environment
meillo@35 1204 matches good here.
meillo@35 1205 There was the question, how anyone could want anything less than the
meillo@35 1206 best program for the job.
meillo@45 1207 But as personal preferences matter, it is probably more important to ask:
meillo@35 1208 How could anyone want something else than his preferred program for the job?
meillo@36 1209 .PP
meillo@45 1210 Usually users want one program for a specific job.
meillo@35 1211 Hence, whenever the task is, for instance, downloading,
meillo@45 1212 the same download manager should be used.
meillo@35 1213 More advanced users might want to have this download manager in this
meillo@35 1214 situation and that one in that situation.
meillo@35 1215 They should be able to configure it this way.
meillo@35 1216 With uzbl, one can use any download manager the user wants.
meillo@36 1217 To switch to a different one, only one line in a small handler script
meillo@35 1218 needs to be changed.
meillo@45 1219 Alternatively it would be possible to query the program to use by
meillo@45 1220 reading a global file or an environment variable, in the handler script.
meillo@36 1221 .PP
meillo@45 1222 Uzbl does neither have its own download manager nor depends on a
meillo@45 1223 specific one, hence uzbl's browsing abilities will not be lowered by having
meillo@35 1224 a bad download manager.
meillo@45 1225 Uzbl's download capabilities will be just as good as the ones of the best
meillo@36 1226 download manager available on the system.
meillo@38 1227 Of course, this applies to all of the other supplementary tools, too.
meillo@32 1228
meillo@32 1229 .PP
meillo@36 1230 .B "Use software leverage to your advantage" .
meillo@36 1231 Uzbl is designed to be extended by external tools.
meillo@36 1232 These external tools are usually wrapped by small handler shell scripts.
meillo@36 1233 Shell scripts are the glue in this approach.
meillo@36 1234 They make the various parts fit together.
meillo@36 1235 .PP
meillo@45 1236 The history mechanism of uzbl shall be presented as an example.
meillo@36 1237 Uzbl is configured to spawn a script to append an entry to the history
meillo@36 1238 whenever the event of a fully loaded page occurs.
meillo@45 1239 The script to append the entry to the history is not much more than:
meillo@36 1240 .DS
meillo@36 1241 #!/bin/sh
meillo@36 1242 file=/path/to/uzbl-history
meillo@36 1243 echo `date +'%Y-%m-%d %H:%M:%S'`" $6 $7" >> $file
meillo@36 1244 .DE
meillo@36 1245 \f(CW$6\fP and \f(CW$7\fP expand to the \s-1URL\s0 and the page title.
meillo@45 1246 .PP
meillo@45 1247 For loading an entry, a key is bound to spawn a load-from-history script.
meillo@36 1248 The script reverses the history to have newer entries first,
meillo@45 1249 then displays \fIdmenu\fP to let the user select an item,
meillo@36 1250 and afterwards writes the selected \s-1URL\s0 into uzbl's command input pipe.
meillo@45 1251 With error checking and corner case handling removed,
meillo@45 1252 the script looks like this:
meillo@36 1253 .DS
meillo@36 1254 #!/bin/sh
meillo@36 1255 file=/path/to/uzbl-history
meillo@36 1256 goto=`tac $file | dmenu | cut -d' ' -f 3`
meillo@36 1257 echo "uri $goto" > $4
meillo@36 1258 .DE
meillo@36 1259 \f(CW$4\fP expands to the path of the command input pipe of the current
meillo@36 1260 uzbl instance.
meillo@32 1261
meillo@32 1262 .PP
meillo@33 1263 .B "Avoid captive user interfaces" .
meillo@36 1264 One could say, that uzbl, to a large extend, actually \fIis\fP
meillo@36 1265 a captive user interface.
meillo@37 1266 But the difference to most other web browsers is, that uzbl is only
meillo@45 1267 the captive user interface frontend (and the core of the backend).
meillo@38 1268 Many parts of the backend are independent of uzbl.
meillo@45 1269 Some are distributed with uzbl, for some external programs,
meillo@45 1270 handler scripts are distributed,
meillo@45 1271 but arbitrary additional functionality can be added if desired.
meillo@37 1272 .PP
meillo@37 1273 The frontend is captive \(en that is true.
meillo@37 1274 This is okay for the task of browsing the web, as this task is only relevant
meillo@37 1275 for humans.
meillo@37 1276 Automated programs would \fIcrawl\fP the web.
meillo@37 1277 That means, they read the source directly.
meillo@37 1278 The source includes all the semantics.
meillo@37 1279 The graphical representation is just for humans to transfer the semantics
meillo@37 1280 more intuitively.
meillo@32 1281
meillo@32 1282 .PP
meillo@33 1283 .B "Make every program a filter" .
meillo@37 1284 Graphical web browsers are almost dead ends in the chain of information flow.
meillo@37 1285 Thus it is difficult to see what graphical web browsers should filter.
meillo@45 1286 Graphical web browsers exist almost only to be interactively used by humans.
meillo@37 1287 The only case when one might want to automate the rendering function is
meillo@37 1288 to generate images of rendered webpages.
meillo@37 1289
meillo@37 1290 .PP
meillo@37 1291 .B "Small is beautiful"
meillo@45 1292 is not easy to apply to a web browser, because modern web technology
meillo@45 1293 is very complex, hence the rendering task is very complex.
meillo@45 1294 Modern web browsers have to consist of many thousand lines of code,
meillo@37 1295 unfortunately.
meillo@37 1296 Using the toolchest approach and wrappers can split the browser into
meillo@37 1297 several small parts, tough.
meillo@37 1298 .PP
meillo@45 1299 As of March 2010, uzbl-core consists of about 3\,500 lines of C code.
meillo@37 1300 The distribution includes another 3\,500 lines of Shell and Python code,
meillo@37 1301 which are the handler scripts and plugins like a modal interface.
meillo@38 1302 Further more, uzbl uses functionality of external tools like
meillo@38 1303 \fIwget\fP and \fInetcat\fP.
meillo@37 1304 Up to this point, uzbl looks pretty neat and small.
meillo@38 1305 The ugly part of uzbl is the web content renderer, webkit.
meillo@37 1306 Webkit consists of roughly 400\,000 (!) lines of code.
meillo@38 1307 Unfortunately, small web render engines are not possible anymore
meillo@38 1308 because of the modern web.
meillo@35 1309
meillo@35 1310 .PP
meillo@35 1311 .B "Build a prototype as soon as possible" .
meillo@35 1312 Plaetinck made his code public, right from the beginning.
meillo@38 1313 Discussion and development was, and still is, open to everyone interested.
meillo@38 1314 Development versions of uzbl can be obtained very simply from the code
meillo@38 1315 repository.
meillo@38 1316 Within the first year of uzbl's existence, a new version was released
meillo@35 1317 more often than once a month.
meillo@38 1318 Different forks and branches arose.
meillo@45 1319 They introduced new features, which were tested for suitability
meillo@45 1320 for the main branch.
meillo@35 1321 The experiences of using prototypes influenced further development.
meillo@35 1322 Actually, all development was community driven.
meillo@38 1323 Plaetinck says, three months after uzbl's birth:
meillo@35 1324 ``Right now I hardly code anything myself for Uzbl.
meillo@35 1325 I just merge in other people's code, ponder a lot, and lead the discussions.''
meillo@35 1326 .[
meillo@44 1327 lwn
meillo@44 1328 uzbl
meillo@35 1329 .]
meillo@32 1330
meillo@0 1331
meillo@48 1332 .H 2 "Problems
meillo@0 1333 .LP
meillo@38 1334 Similar to \s-1MH\s0, uzbl, too suffers from being different.
meillo@38 1335 It is sad, but people use what they know.
meillo@38 1336 Fortunately, uzbl's user interface can look and feel very much the
meillo@38 1337 same as the one of the well known web browsers,
meillo@38 1338 hiding the internal differences.
meillo@38 1339 But uzbl has to provide this similar look and feel to be accepted
meillo@38 1340 as a ``normal'' browser by ``normal'' users.
meillo@37 1341 .PP
meillo@45 1342 Though, the more important problem is the modern web.
meillo@38 1343 The modern web is simply broken.
meillo@38 1344 It has state in a state-less protocol,
meillo@38 1345 it misuses technologies,
meillo@38 1346 and it is helplessly overloaded.
meillo@38 1347 The result are web content render engines that must consist
meillo@38 1348 of hundreds of thousands lines of code.
meillo@38 1349 They also must combine and integrate many different technologies,
meillo@45 1350 only to make our modern web accessible.
meillo@38 1351 Website to image converter are hardly possible to run without
meillo@38 1352 human interaction because of state in sessions, impossible
meillo@38 1353 deep-linking, and unautomatable technologies.
meillo@37 1354 .PP
meillo@38 1355 The web was misused to provide all kinds of imaginable wishes.
meillo@38 1356 Now web browsers, and eventually the users, suffer from it.
meillo@37 1357
meillo@8 1358
meillo@48 1359 .H 2 "Summary uzbl
meillo@32 1360 .LP
meillo@38 1361 ``Uzbl is a browser that adheres to the Unix Philosophy'',
meillo@38 1362 that is how uzbl is seen by its authors.
meillo@38 1363 Indeed, uzbl follows the Unix Philosophy in many ways.
meillo@38 1364 It consists of independent parts that work together,
meillo@45 1365 while its core is mainly a mediator which glues the parts together.
meillo@38 1366 .PP
meillo@38 1367 Software leverage can excellently be seen in uzbl.
meillo@45 1368 External tools are used, independent tasks are separated
meillo@45 1369 in independent parts and glued together with small handler scripts.
meillo@38 1370 .PP
meillo@38 1371 As uzbl, more or less, consists of a set of tools and a bit
meillo@38 1372 of glue, anyone can put the parts together and expand it
meillo@38 1373 in any desired way.
meillo@38 1374 Uzbl is very flexible and customizable.
meillo@38 1375 These properties make it valuable for advanced users,
meillo@38 1376 but may keep novice users from using it.
meillo@38 1377 .PP
meillo@45 1378 But uzbl's main problem is the modern web, that makes it hard
meillo@38 1379 to design a sane web browser.
meillo@38 1380 Despite this bad situation, uzbl does a fairly good job.
meillo@32 1381
meillo@8 1382
meillo@48 1383 .H 1 "Final thoughts
meillo@0 1384
meillo@48 1385 .H 2 "Quick summary
meillo@0 1386 .LP
meillo@0 1387 good design
meillo@0 1388 .LP
meillo@0 1389 unix phil
meillo@0 1390 .LP
meillo@0 1391 case studies
meillo@0 1392
meillo@48 1393 .H 2 "Why people should choose
meillo@0 1394 .LP
meillo@0 1395 Make the right choice!
meillo@0 1396
meillo@47 1397
meillo@47 1398 .bp
meillo@47 1399 .TL
meillo@47 1400 References
meillo@47 1401 .LP
meillo@47 1402 .XS
meillo@47 1403 .sp .5v
meillo@47 1404 .B
meillo@47 1405 References
meillo@47 1406 .XE
meillo@47 1407 .ev r
meillo@42 1408 .nr PS -1
meillo@42 1409 .nr VS -1
meillo@0 1410 .[
meillo@0 1411 $LIST$
meillo@0 1412 .]
meillo@47 1413 .nr PS +1
meillo@47 1414 .nr VS +1
meillo@47 1415 .ev
meillo@47 1416
meillo@42 1417 .bp
meillo@47 1418 .TL
meillo@47 1419 Table of Contents
meillo@47 1420 .LP
meillo@47 1421 .PX no