annotate @ 13:c665b9e7bf8a

more about portability and reuse (ch03)
date Sun, 21 Feb 2010 14:52:53 +0100
parents 018d5f9a2993
children 59305c854751
rev   line source
meillo@2 1 .\".if n .pl 1000i
meillo@0 2 .de XX
meillo@0 3 .pl 1v
meillo@0 4 ..
meillo@0 5 .em XX
meillo@1 6 .\".nr PI 0
meillo@1 7 .\".if t .nr PD .5v
meillo@1 8 .\".if n .nr PD 1v
meillo@0 9 .nr lu 0
meillo@0 10 .de CW
meillo@0 11 .nr PQ \\n(.f
meillo@0 12 .if t .ft CW
meillo@0 13 .ie \\$1 .if n .ul 999
meillo@0 14 .el .if n .ul 1
meillo@0 15 .if t .if !\\$1 \&\\$1\f\\n(PQ\\$2
meillo@0 16 .if n .if \\n(.$=1 \&\\$1
meillo@0 17 .if n .if \\n(.$>1 \&\\$1\c
meillo@0 18 .if n .if \\n(.$>1 \&\\$2
meillo@0 19 ..
meillo@0 20 .ds [. \ [
meillo@0 21 .ds .] ]
meillo@1 22 .\"----------------------------------------
meillo@0 23 .TL
meillo@6 24 Why the Unix Philosophy still matters
meillo@0 25 .AU
meillo@0 26 markus schnalke <>
meillo@0 27 .AB
meillo@1 28 .ti \n(.iu
meillo@2 29 This paper discusses the importance of the Unix Philosophy in software design.
meillo@0 30 Today, few software designers are aware of these concepts,
meillo@3 31 and thus most modern software is limited and does not make use of software leverage.
meillo@0 32 Knowing and following the tenets of the Unix Philosophy makes software more valuable.
meillo@0 33 .AE
meillo@0 34
meillo@10 35 .\".if t .2C
meillo@2 36
meillo@2 37 .FS
meillo@2 38 .ps -1
meillo@2 39 This paper was prepared for the seminar ``Software Analysis'' at University Ulm.
meillo@2 40 Mentor was professor Schweiggert. 2010-02-05
meillo@2 41 .br
meillo@2 42 You may get this document from my website
meillo@2 43 .CW \s-1
meillo@2 44 .FE
meillo@2 45
meillo@0 46 .NH 1
meillo@0 47 Introduction
meillo@0 48 .LP
meillo@0 49 Building a software is a process from an idea of the purpose of the software
meillo@3 50 to its release.
meillo@0 51 No matter \fIhow\fP the process is run, two things are common:
meillo@0 52 the initial idea and the release.
meillo@9 53 The process in between can be of any shape.
meillo@9 54 The the maintenance work after the release is ignored for the moment.
meillo@1 55 .PP
meillo@0 56 The process of building splits mainly in two parts:
meillo@0 57 the planning of what and how to build, and implementing the plan by writing code.
meillo@3 58 This paper focuses on the planning part \(en the designing of the software.
meillo@3 59 .PP
meillo@3 60 Software design is the plan of how the internals and externals of the software should look like,
meillo@3 61 based on the requirements.
meillo@9 62 This paper discusses the recommendations of the Unix Philosophy about software design.
meillo@3 63 .PP
meillo@3 64 The here discussed ideas can get applied by any development process.
meillo@9 65 The Unix Philosophy does recommend how the software development process should look like,
meillo@3 66 but this shall not be of matter here.
meillo@0 67 Similar, the question of how to write the code is out of focus.
meillo@1 68 .PP
meillo@3 69 The name ``Unix Philosophy'' was already mentioned several times, but it was not explained yet.
meillo@1 70 The Unix Philosophy is the essence of how the Unix operating system and its toolchest was designed.
meillo@3 71 It is no limited set of rules, but what people see to be common to typical Unix software.
meillo@1 72 Several people stated their view on the Unix Philosophy.
meillo@1 73 Best known are:
meillo@1 74 .IP \(bu
meillo@1 75 Doug McIlroy's summary: ``Write programs that do one thing and do it well.''
meillo@1 76 .[
meillo@1 77 %A M. D. McIlroy
meillo@1 78 %A E. N. Pinson
meillo@1 79 %A B. A. Taque
meillo@1 80 %T UNIX Time-Sharing System Forward
meillo@1 81 %J The Bell System Technical Journal
meillo@1 82 %D 1978
meillo@1 83 %V 57
meillo@1 84 %N 6
meillo@1 85 %P 1902
meillo@1 86 .]
meillo@1 87 .IP \(bu
meillo@1 88 Mike Gancarz' book ``The UNIX Philosophy''.
meillo@1 89 .[
meillo@1 90 %A Mike Gancarz
meillo@1 91 %T The UNIX Philosophy
meillo@1 92 %D 1995
meillo@1 93 %I Digital Press
meillo@1 94 .]
meillo@1 95 .IP \(bu
meillo@1 96 Eric S. Raymond's book ``The Art of UNIX Programming''.
meillo@1 97 .[
meillo@1 98 %A Eric S. Raymond
meillo@1 99 %T The Art of UNIX Programming
meillo@1 100 %D 2003
meillo@1 101 %I Addison-Wesley
meillo@2 102 %O .CW \s-1
meillo@1 103 .]
meillo@0 104 .LP
meillo@1 105 These different views on the Unix Philosophy have much in common.
meillo@3 106 Especially, the main concepts are similar for all of them.
meillo@1 107 But there are also points on which they differ.
meillo@1 108 This only underlines what the Unix Philosophy is:
meillo@1 109 A retrospective view on the main concepts of Unix software;
meillo@9 110 especially those that were successful and unique to Unix.
meillo@6 111 .\" really?
meillo@1 112 .PP
meillo@1 113 Before we will have a look at concrete concepts,
meillo@1 114 we discuss why software design is important
meillo@1 115 and what problems bad design introduces.
meillo@0 116
meillo@0 117
meillo@0 118 .NH 1
meillo@6 119 Importance of software design in general
meillo@0 120 .LP
meillo@2 121 Why should we design software at all?
meillo@6 122 It is general knowledge, that even a bad plan is better than no plan.
meillo@6 123 Ignoring software design is programming without a plan.
meillo@6 124 This will lead pretty sure to horrible results.
meillo@2 125 .PP
meillo@6 126 The design of a software is its internal and external shape.
meillo@6 127 The design talked about here has nothing to do with visual appearance.
meillo@6 128 If we see a program as a car, then its color is of no matter.
meillo@6 129 Its design would be the car's size, its shape, the number and position of doors,
meillo@6 130 the ratio of passenger and cargo transport, and so forth.
meillo@2 131 .PP
meillo@6 132 A software's design is about quality properties.
meillo@6 133 Each of the cars may be able to drive from A to B,
meillo@6 134 but it depends on its properties whether it is a good car for passenger transport or not.
meillo@6 135 It also depends on its properties if it is a good choice for a rough mountain area.
meillo@2 136 .PP
meillo@6 137 Requirements to a software are twofold: functional and non-functional.
meillo@6 138 Functional requirements are easier to define and to verify.
meillo@6 139 They are directly the software's functions.
meillo@6 140 Functional requirements are the reason why software gets written.
meillo@6 141 Someone has a problem and needs a tool to solve it.
meillo@6 142 Being able to solve the problem is the main functional requirement.
meillo@6 143 It is the driving force behind all programming effort.
meillo@2 144 .PP
meillo@6 145 On the other hand, there are also non-functional requirements.
meillo@6 146 They are called \fIquality\fP requirements, too.
meillo@6 147 The quality of a software is about properties that are not directly related to
meillo@6 148 the software's basic functions.
meillo@6 149 Quality aspects are about the properties that are overlooked at first sight.
meillo@2 150 .PP
meillo@6 151 Quality is of few matter when the software gets initially built,
meillo@9 152 but it will be of matter in usage and maintenance of the software.
meillo@6 153 A short-sighted might see in developing a software mainly building something up.
meillo@6 154 Reality shows, that building the software the first time is only a small amount
meillo@6 155 of the overall work.
meillo@9 156 Bug fixing, extending, rebuilding of parts \(en short: maintenance work \(en
meillo@6 157 does soon take over the major part of the time spent on a software.
meillo@6 158 Not to forget the usage of the software.
meillo@6 159 These processes are highly influenced by the software's quality.
meillo@6 160 Thus, quality should never be neglected.
meillo@6 161 The problem is that you hardly ``stumble over'' bad quality during the first build,
meillo@6 162 but this is the time when you should care about good quality most.
meillo@6 163 .PP
meillo@6 164 Software design is not about the basic function of a software;
meillo@6 165 this requirement will get satisfied anyway, as it is the main driving force behind the development.
meillo@6 166 Software design is about quality aspects of the software.
meillo@6 167 Good design will lead to good quality, bad design to bad quality.
meillo@6 168 The primary functions of the software will be affected modestly by bad quality,
meillo@6 169 but good quality can provide a lot of additional gain from the software,
meillo@6 170 even at places where one never expected it.
meillo@6 171 .PP
meillo@6 172 The ISO/IEC 9126-1 standard, part 1,
meillo@6 173 .[
meillo@9 174 %I International Organization for Standardization
meillo@6 175 %T ISO Standard 9126: Software Engineering \(en Product Quality, part 1
meillo@6 176 %C Geneve
meillo@6 177 %D 2001
meillo@6 178 .]
meillo@6 179 defines the quality model as consisting out of:
meillo@6 180 .IP \(bu
meillo@6 181 .I Functionality
meillo@6 182 (suitability, accuracy, inter\%operability, security)
meillo@6 183 .IP \(bu
meillo@6 184 .I Reliability
meillo@6 185 (maturity, fault tolerance, recoverability)
meillo@6 186 .IP \(bu
meillo@6 187 .I Usability
meillo@6 188 (understandability, learnability, operability, attractiveness)
meillo@6 189 .IP \(bu
meillo@6 190 .I Efficiency
meillo@9 191 (time behavior, resource utilization)
meillo@6 192 .IP \(bu
meillo@6 193 .I Maintainability
meillo@6 194 (analysability, changeability, stability, testability)
meillo@6 195 .IP \(bu
meillo@6 196 .I Portability
meillo@6 197 (adaptability, installability, co-existence, replaceability)
meillo@6 198 .LP
meillo@6 199 These goals are parts of a software's design.
meillo@6 200 Good design can give these properties to a software,
meillo@6 201 bad designed software will miss them.
meillo@7 202 .PP
meillo@7 203 One further goal of software design is consistency.
meillo@7 204 Consistency eases understanding, working on, and using things.
meillo@7 205 Consistent internals and consistent interfaces to the outside can be provided by good design.
meillo@7 206 .PP
meillo@7 207 We should design software because good design avoids many problems during a software's lifetime.
meillo@7 208 And we should design software because good design can offer much gain,
meillo@7 209 that can be unrelated to the software main intend.
meillo@7 210 Indeed, we should spend much effort into good design to make the software more valuable.
meillo@7 211 The Unix Philosophy shows how to design software well.
meillo@7 212 It offers guidelines to achieve good quality and high gain for the effort spent.
meillo@0 213
meillo@0 214
meillo@0 215 .NH 1
meillo@0 216 The Unix Philosophy
meillo@4 217 .LP
meillo@4 218 The origins of the Unix Philosophy were already introduced.
meillo@8 219 This chapter explains the philosophy, oriented on Gancarz,
meillo@8 220 and shows concrete examples of its application.
meillo@5 221
meillo@5 222 .SH
meillo@4 223 Examples
meillo@4 224 .LP
meillo@4 225 Following are some examples to demonstrate how applied Unix Philosophy feels like.
meillo@4 226 Knowledge of using the Unix shell is assumed.
meillo@4 227 .PP
meillo@4 228 Counting the number of files in the current directory:
meillo@9 229 .DS I 2n
meillo@4 230 .CW
meillo@9 231 .ps -1
meillo@4 232 ls | wc -l
meillo@4 233 .DE
meillo@4 234 The
meillo@4 235 .CW ls
meillo@4 236 command lists all files in the current directory, one per line,
meillo@4 237 and
meillo@4 238 .CW "wc -l
meillo@8 239 counts the number of lines.
meillo@4 240 .PP
meillo@8 241 Counting the number of files that do not contain ``foo'' in their name:
meillo@9 242 .DS I 2n
meillo@4 243 .CW
meillo@9 244 .ps -1
meillo@4 245 ls | grep -v foo | wc -l
meillo@4 246 .DE
meillo@4 247 Here, the list of files is filtered by
meillo@4 248 .CW grep
meillo@4 249 to remove all that contain ``foo''.
meillo@4 250 The rest is the same as in the previous example.
meillo@4 251 .PP
meillo@4 252 Finding the five largest entries in the current directory.
meillo@9 253 .DS I 2n
meillo@4 254 .CW
meillo@9 255 .ps -1
meillo@4 256 du -s * | sort -nr | sed 5q
meillo@4 257 .DE
meillo@4 258 .CW "du -s *
meillo@4 259 returns the recursively summed sizes of all files
meillo@8 260 \(en no matter if they are regular files or directories.
meillo@4 261 .CW "sort -nr
meillo@4 262 sorts the list numerically in reverse order.
meillo@4 263 Finally,
meillo@4 264 .CW "sed 5q
meillo@4 265 quits after it has printed the fifth line.
meillo@4 266 .PP
meillo@4 267 The presented command lines are examples of what Unix people would use
meillo@4 268 to get the desired output.
meillo@4 269 There are also other ways to get the same output.
meillo@4 270 It's a user's decision which way to go.
meillo@5 271
meillo@5 272 .SH
meillo@4 273 Pipes
meillo@4 274 .LP
meillo@8 275 The examples show that many tasks on a Unix system
meillo@4 276 are accomplished by combining several small programs.
meillo@4 277 The connection between the single programs is denoted by the pipe operator `|'.
meillo@4 278 .PP
meillo@4 279 Pipes, and their extensive and easy use, are one of the great
meillo@4 280 achievements of the Unix system.
meillo@4 281 Pipes between programs have been possible in earlier operating systems,
meillo@4 282 but it has never been a so central part of the concept.
meillo@4 283 When, in the early seventies, Doug McIlroy introduced pipes for the
meillo@4 284 Unix system,
meillo@4 285 ``it was this concept and notation for linking several programs together
meillo@4 286 that transformed Unix from a basic file-sharing system to an entirely new way of computing.''
meillo@4 287 .[
meillo@4 288 %T Unix: An Oral History
meillo@5 289 %O .CW \s-1
meillo@4 290 .]
meillo@4 291 .PP
meillo@4 292 Being able to specify pipelines in an easy way is,
meillo@4 293 however, not enough by itself.
meillo@5 294 It is only one half.
meillo@4 295 The other is the design of the programs that are used in the pipeline.
meillo@8 296 They have to interfaces that allows them to be used in such a way.
meillo@5 297
meillo@5 298 .SH
meillo@5 299 Interface architecture
meillo@5 300 .LP
meillo@11 301 Unix is, first of all, simple \(en Everything is a file.
meillo@5 302 Files are sequences of bytes, without any special structure.
meillo@5 303 Programs should be filters, which read a stream of bytes from ``standard input'' (stdin)
meillo@5 304 and write a stream of bytes to ``standard output'' (stdout).
meillo@5 305 .PP
meillo@8 306 If the files \fIare\fP sequences of bytes,
meillo@8 307 and the programs \fIare\fP filters on byte streams,
meillo@11 308 then there is exactly one standardized data interface.
meillo@5 309 Thus it is possible to combine them in any desired way.
meillo@5 310 .PP
meillo@5 311 Even a handful of small programs will yield a large set of combinations,
meillo@5 312 and thus a large set of different functions.
meillo@5 313 This is leverage!
meillo@5 314 If the programs are orthogonal to each other \(en the best case \(en
meillo@5 315 then the set of different functions is greatest.
meillo@5 316 .PP
meillo@11 317 Programs might also have a separate control interface,
meillo@11 318 besides their data interface.
meillo@11 319 The control interface is often called ``user interface'',
meillo@11 320 because it is usually designed to be used by humans.
meillo@11 321 The Unix Philosophy discourages to assume the user to be human.
meillo@11 322 Interactive use of software is slow use of software,
meillo@11 323 because the program waits for user input most of the time.
meillo@11 324 Interactive software requires the user to be in front of the computer
meillo@11 325 all the time.
meillo@11 326 Interactive software occupy the user's attention while they are running.
meillo@11 327 .PP
meillo@11 328 Now we come back to the idea of using several small programs, combined,
meillo@11 329 to have a more specific function.
meillo@11 330 If these single tools would all be interactive,
meillo@11 331 how would the user control them?
meillo@11 332 It is not only a problem to control several programs at once if they run at the same time,
meillo@11 333 it also very inefficient to have to control each of the single programs
meillo@11 334 that are intended to work as one large program.
meillo@11 335 Hence, the Unix Philosophy discourages programs to demand interactive use.
meillo@11 336 The behavior of programs should be defined at invocation.
meillo@11 337 This is done by specifying arguments (``command line switches'') to the program call.
meillo@11 338 Gancarz discusses this topic as ``avoid captive user interfaces''.
meillo@11 339 .[
meillo@11 340 %A Mike Gancarz
meillo@11 341 %T The UNIX Philosophy
meillo@11 342 %I Digital Press
meillo@11 343 %D 1995
meillo@11 344 %P 88 ff.
meillo@11 345 .]
meillo@11 346 .PP
meillo@11 347 Non-interactive use is, during development, also an advantage for testing.
meillo@11 348 Testing of interactive programs is much more complicated,
meillo@11 349 than testing of non-interactive programs.
meillo@5 350
meillo@5 351 .SH
meillo@8 352 The toolchest approach
meillo@5 353 .LP
meillo@5 354 A toolchest is a set of tools.
meillo@5 355 Instead of having one big tool for all tasks, one has many small tools,
meillo@5 356 each for one task.
meillo@5 357 Difficult tasks are solved by combining several of the small, simple tools.
meillo@5 358 .PP
meillo@11 359 The Unix toolchest \fIis\fP a set of small, (mostly) non-interactive programs
meillo@11 360 that are filters on byte streams.
meillo@11 361 They are, to a large extend, unrelated in their function.
meillo@11 362 Hence, the Unix toolchest provides a large set of functions
meillo@11 363 that can be accessed by combining the programs in the desired way.
meillo@11 364 .PP
meillo@11 365 There are also advantages for developing small toolchest programs.
meillo@5 366 It is easier and less error-prone to write small programs.
meillo@5 367 It is also easier and less error-prone to write a large set of small programs,
meillo@5 368 than to write one large program with all the functionality included.
meillo@5 369 If the small programs are combinable, then they offer even a larger set
meillo@5 370 of functions than the single large program.
meillo@5 371 Hence, one gets two advantages out of writing small, combinable programs.
meillo@5 372 .PP
meillo@5 373 There are two drawbacks of the toolchest approach.
meillo@8 374 First, one simple, standardized, unidirectional interface has to be sufficient.
meillo@5 375 If one feels the need for more ``logic'' than a stream of bytes,
meillo@8 376 then a different approach might be of need.
meillo@13 377 But it is also possible, that he just can not imagine a design where
meillo@8 378 a stream of bytes is sufficient.
meillo@8 379 By becoming more familiar with the ``Unix style of thinking'',
meillo@8 380 developers will more often and easier find simple designs where
meillo@8 381 a stream of bytes is a sufficient interface.
meillo@8 382 .PP
meillo@8 383 The second drawback of a toolchest affects the users.
meillo@5 384 A toolchest is often more difficult to use for novices.
meillo@9 385 It is necessary to become familiar with each of the tools,
meillo@5 386 to be able to use the right one in a given situation.
meillo@9 387 Additionally, one needs to combine the tools in a senseful way on its own.
meillo@9 388 This is like a sharp knife \(en it is a powerful tool in the hand of a master,
meillo@5 389 but of no good value in the hand of an unskilled.
meillo@5 390 .PP
meillo@8 391 However, learning single, small tool of the toolchest is easier than
meillo@8 392 learning a complex tool.
meillo@8 393 The user will have a basic understanding of a yet unknown tool,
meillo@8 394 if the several tools of the toolchest have a common style.
meillo@8 395 He will be able to transfer knowledge over one tool to another.
meillo@8 396 .PP
meillo@8 397 Moreover, the second drawback can be removed easily by adding wrappers
meillo@8 398 around the single tools.
meillo@5 399 Novice users do not need to learn several tools if a professional wraps
meillo@8 400 the single commands into a more high-level script.
meillo@5 401 Note that the wrapper script still calls the small tools;
meillo@5 402 the wrapper script is just like a skin around.
meillo@8 403 No complexity is added this way,
meillo@8 404 but new programs can get created out of existing one with very low effort.
meillo@5 405 .PP
meillo@5 406 A wrapper script for finding the five largest entries in the current directory
meillo@5 407 could look like this:
meillo@9 408 .DS I 2n
meillo@5 409 .CW
meillo@9 410 .ps -1
meillo@5 411 #!/bin/sh
meillo@5 412 du -s * | sort -nr | sed 5q
meillo@5 413 .DE
meillo@5 414 The script itself is just a text file that calls the command line
meillo@5 415 a professional user would type in directly.
meillo@8 416 Making the program flexible on the number of entries it prints,
meillo@8 417 is easily possible:
meillo@9 418 .DS I 2n
meillo@8 419 .CW
meillo@9 420 .ps -1
meillo@8 421 #!/bin/sh
meillo@8 422 num=5
meillo@8 423 [ $# -eq 1 ] && num="$1"
meillo@8 424 du -sh * | sort -nr | sed "${num}q"
meillo@8 425 .DE
meillo@8 426 This script acts like the one before, when called without an argument.
meillo@8 427 But one can also specify a numerical argument to define the number of lines to print.
meillo@5 428
meillo@8 429 .SH
meillo@8 430 A powerful shell
meillo@8 431 .LP
meillo@10 432 It was already said, that the Unix shell provides the possibility to
meillo@10 433 combine small programs into large ones easily.
meillo@10 434 A powerful shell is a great feature in other ways, too.
meillo@8 435 .PP
meillo@10 436 For instance by including a scripting language.
meillo@10 437 The control statements are build into the shell.
meillo@8 438 The functions, however, are the normal programs, everyone can use on the system.
meillo@10 439 Thus, the programs are known, so learning to program in the shell is easy.
meillo@8 440 Using normal programs as functions in the shell programming language
meillo@10 441 is only possible because they are small and combinable tools in a toolchest style.
meillo@8 442 .PP
meillo@8 443 The Unix shell encourages to write small scripts out of other programs,
meillo@8 444 because it is so easy to do.
meillo@8 445 This is a great step towards automation.
meillo@8 446 It is wonderful if the effort to automate a task equals the effort
meillo@8 447 it takes to do it the second time by hand.
meillo@8 448 If it is so, then the user will be happy to automate everything he does more than once.
meillo@8 449 .PP
meillo@8 450 Small programs that do one job well, standardized interfaces between them,
meillo@8 451 a mechanism to combine parts to larger parts, and an easy way to automate tasks,
meillo@8 452 this will inevitably produce software leverage.
meillo@8 453 Getting multiple times the benefit of an investment is a great offer.
meillo@10 454 .PP
meillo@10 455 The shell also encourages rapid prototyping.
meillo@10 456 Many well known programs started as quickly hacked shell scripts,
meillo@10 457 and turned into ``real'' programs, written in C, later.
meillo@10 458 Building a prototype first is a way to avoid the biggest problems
meillo@10 459 in application development.
meillo@10 460 Fred Brooks writes in ``No Silver Bullet'':
meillo@10 461 .[
meillo@10 462 %A Frederick P. Brooks, Jr.
meillo@10 463 %T No Silver Bullet: Essence and Accidents of Software Engineering
meillo@10 464 %B Information Processing 1986, the Proceedings of the IFIP Tenth World Computing Conference
meillo@10 465 %E H.-J. Kugler
meillo@10 466 %D 1986
meillo@10 467 %P 1069\(en1076
meillo@10 468 %I Elsevier Science B.V.
meillo@10 469 %C Amsterdam, The Netherlands
meillo@10 470 .]
meillo@10 471 .QP
meillo@10 472 The hardest single part of building a software system is deciding precisely what to build.
meillo@10 473 No other part of the conceptual work is so difficult as establishing the detailed
meillo@10 474 technical requirements, [...].
meillo@10 475 No other part of the work so cripples the resulting system if done wrong.
meillo@10 476 No other part is more difficult to rectify later.
meillo@10 477 .PP
meillo@10 478 Writing a prototype is a great method to become familiar with the requirements
meillo@10 479 and to actually run into real problems.
meillo@10 480 Today, prototyping is often seen as a first step in building a software.
meillo@10 481 This is, of course, good.
meillo@10 482 However, the Unix Philosophy has an \fIadditional\fP perspective on prototyping:
meillo@10 483 After having built the prototype, one might notice, that the prototype is already
meillo@10 484 \fIgood enough\fP.
meillo@10 485 Hence, no reimplementation, in a more sophisticated programming language, might be of need,
meillo@10 486 for the moment.
meillo@10 487 Maybe later, it might be neccessary to rewrite the software, but not now.
meillo@10 488 .PP
meillo@10 489 By delaying further work, one keeps the flexibility to react easily on
meillo@10 490 changing requirements.
meillo@10 491 Software parts that are not written will not miss the requirements.
meillo@10 492
meillo@10 493 .SH
meillo@10 494 Worse is better
meillo@10 495 .LP
meillo@10 496 The Unix Philosophy aims for the 80% solution;
meillo@10 497 others call it the ``Worse is better'' approach.
meillo@10 498 .PP
meillo@10 499 First, practical experience shows, that it is almost never possible to define the
meillo@10 500 requirements completely and correctly the first time.
meillo@10 501 Hence one should not try to; it will fail anyway.
meillo@10 502 Second, practical experience shows, that requirements change during time.
meillo@10 503 Hence it is best to delay requirement-based design decisions as long as possible.
meillo@10 504 Also, the software should be small and flexible as long as possible
meillo@10 505 to react on changing requirements.
meillo@10 506 Shell scripts, for example, are more easily adjusted as C programs.
meillo@10 507 Third, practical experience shows, that maintenance is hard work.
meillo@10 508 Hence, one should keep the amount of software as small as possible;
meillo@10 509 it should just fulfill the \fIcurrent\fP requirements.
meillo@10 510 Software parts that will be written later, do not need maintenance now.
meillo@10 511 .PP
meillo@10 512 Starting with a prototype in a scripting language has several advantages:
meillo@10 513 .IP \(bu
meillo@10 514 As the initial effort is low, one will likely start right away.
meillo@10 515 .IP \(bu
meillo@10 516 As working parts are available soon, the real requirements can get identified soon.
meillo@10 517 .IP \(bu
meillo@10 518 When a software is usable, it gets used, and thus tested.
meillo@10 519 Hence problems will be found at early stages of the development.
meillo@10 520 .IP \(bu
meillo@10 521 The prototype might be enough for the moment,
meillo@10 522 thus further work on the software can be delayed to a time
meillo@10 523 when one knows better about the requirements and problems,
meillo@10 524 than now.
meillo@10 525 .IP \(bu
meillo@10 526 Implementing now only the parts that are actually needed now,
meillo@10 527 requires fewer maintenance work.
meillo@10 528 .IP \(bu
meillo@10 529 If the global situation changes so that the software is not needed anymore,
meillo@10 530 then less effort was spent into the project, than it would have be
meillo@10 531 when a different approach had been used.
meillo@10 532
meillo@11 533 .SH
meillo@11 534 Upgrowth and survival of software
meillo@11 535 .LP
meillo@12 536 So far it was talked about \fIwriting\fP or \fIbuilding\fP software.
meillo@13 537 Although these are just verbs, they do imply a specific view on the work process
meillo@13 538 they describe.
meillo@12 539 The better verb, however, is to \fIgrow\fP.
meillo@12 540 .PP
meillo@12 541 Creating software in the sense of the Unix Philosophy is an incremental process.
meillo@12 542 It starts with a first prototype, which evolves as requirements change.
meillo@12 543 A quickly hacked shell script might become a large, sophisticated,
meillo@13 544 compiled program this way.
meillo@13 545 Its lifetime begins with the initial prototype and ends when the software is not used anymore.
meillo@13 546 While being alive it will get extended, rearranged, rebuilt (from scratch).
meillo@12 547 Growing software matches the view that ``software is never finished. It is only released.''
meillo@12 548 .[
meillo@13 549 %O FIXME
meillo@13 550 %A Mike Gancarz
meillo@13 551 %T The UNIX Philosophy
meillo@13 552 %P 26
meillo@12 553 .]
meillo@12 554 .PP
meillo@13 555 Software can be seen as being controlled by evolutionary processes.
meillo@13 556 Successful software is software that is used by many for a long time.
meillo@12 557 This implies that the software is needed, useful, and better than alternatives.
meillo@12 558 Darwin talks about: ``The survival of the fittest.''
meillo@12 559 .[
meillo@13 560 %O FIXME
meillo@13 561 %A Charles Darwin
meillo@12 562 .]
meillo@12 563 Transferred to software: The most successful software, is the fittest,
meillo@12 564 is the one that survives.
meillo@13 565 (This may be at the level of one creature, or at the level of one species.)
meillo@13 566 The fitness of software is affected mainly by four properties:
meillo@12 567 .IP \(bu
meillo@12 568 portability of code
meillo@12 569 .IP \(bu
meillo@12 570 portability of data
meillo@12 571 .IP \(bu
meillo@12 572 range of usability
meillo@12 573 .IP \(bu
meillo@12 574 reuseability of parts
meillo@13 575 .PP
meillo@13 576 Portability of code means, using high-level programming languages,
meillo@13 577 sticking to the standard,
meillo@13 578 and avoiding optimizations that introduce dependencies on specific hardware.
meillo@13 579 Hardware has a much lower lifetime than software.
meillo@13 580 By chaining software to a specific hardware,
meillo@13 581 the software's lifetime gets shortened to that of this hardware.
meillo@13 582 In contrast, software should be easy to port \(en
meillo@13 583 adaption is the key to success.
meillo@13 584 .\" cf. practice of prog: ch08
meillo@13 585 .PP
meillo@13 586 Portability of data is best achieved by avoiding binary representations
meillo@13 587 to store data, because binary representations differ from machine to machine.
meillo@13 588 Textual represenation is favored.
meillo@13 589 Historically, ASCII was the charset of choice.
meillo@13 590 In the future, UTF-8 might be the better choice, however.
meillo@13 591 Important is that it is a plain text representation in a
meillo@13 592 very common charset encoding.
meillo@13 593 Apart from being able to transfer data between machines,
meillo@13 594 readable data has the great advantage, that humans are able
meillo@13 595 to directly edit it with text editors and other tools from the Unix toolchest.
meillo@13 596 .\" gancarz tenet 5
meillo@13 597 .PP
meillo@13 598 A large range of usability ensures good adaption, and thus good survival.
meillo@13 599 It is a special distinction if a software becomes used in fields of action,
meillo@13 600 the original authors did never imagine.
meillo@13 601 Software that solves problems in a general way will likely be used
meillo@13 602 for all kinds of similar problems.
meillo@13 603 Being too specific limits the range of uses.
meillo@13 604 Requirements change through time, thus use cases change or even vanish.
meillo@13 605 A good example in this point is Allman's sendmail.
meillo@13 606 Allman identifies flexibility to be one major reason for sendmail's success:
meillo@13 607 .[
meillo@13 608 %O FIXME
meillo@13 609 %A Allman
meillo@13 610 %T sendmail
meillo@13 611 .]
meillo@13 612 .QP
meillo@13 613 Second, I limited myself to the routing function [...].
meillo@13 614 This was a departure from the dominant thought of the time, [...].
meillo@13 615 .QP
meillo@13 616 Third, the sendmail configuration file was flexible enough to adopt
meillo@13 617 to a rapidly changing world [...].
meillo@12 618 .LP
meillo@13 619 Successful software adopts itself to the changing world.
meillo@13 620 .PP
meillo@13 621 Reuse of parts is even one step further.
meillo@13 622 A software may completely lose its field of action,
meillo@13 623 but parts of which the software is build may be general and independent enough
meillo@13 624 to survive this death.
meillo@13 625 If software is build by combining small independent programs,
meillo@13 626 then there are parts readily available for reuse.
meillo@13 627 Who cares if the large program is a failure,
meillo@13 628 but parts of it become successful instead?
meillo@10 629
meillo@13 630 .SH
meillo@0 631 Results
meillo@0 632 .LP
meillo@0 633 The unix phil is an answer to the sw design question
meillo@0 634 .LP
meillo@0 635 tool chains empower the uses of sw
meillo@0 636
meillo@8 637
meillo@8 638
meillo@0 639 .NH 1
meillo@0 640 Case study: nmh
meillo@0 641
meillo@0 642 .NH 2
meillo@0 643 History
meillo@0 644 .LP
meillo@0 645 MH, nmh.
meillo@0 646 They are old.
meillo@0 647
meillo@0 648 .NH 2
meillo@0 649 Contrasts to similar sw
meillo@0 650 .LP
meillo@0 651 vs. Thunderbird, mutt, mailx, pine
meillo@0 652 .LP
meillo@0 653 flexibility, no redundancy, use the shell
meillo@0 654
meillo@0 655 .NH 2
meillo@0 656 Gains of the design
meillo@0 657 .LP
meillo@0 658
meillo@0 659 .NH 2
meillo@0 660 Problems
meillo@0 661 .LP
meillo@0 662
meillo@8 663
meillo@8 664
meillo@0 665 .NH 1
meillo@0 666 Case study: uzbl
meillo@0 667
meillo@0 668 .NH 2
meillo@0 669 History
meillo@0 670 .LP
meillo@0 671 uzbl is young
meillo@0 672
meillo@0 673 .NH 2
meillo@0 674 Contrasts to similar sw
meillo@0 675 .LP
meillo@0 676 like with nmh
meillo@0 677 .LP
meillo@0 678 addons, plugins, modules
meillo@0 679
meillo@0 680 .NH 2
meillo@0 681 Gains of the design
meillo@0 682 .LP
meillo@0 683
meillo@0 684 .NH 2
meillo@0 685 Problems
meillo@0 686 .LP
meillo@0 687 broken web
meillo@0 688
meillo@8 689
meillo@8 690
meillo@0 691 .NH 1
meillo@0 692 Final thoughts
meillo@0 693
meillo@0 694 .NH 2
meillo@0 695 Quick summary
meillo@0 696 .LP
meillo@0 697 good design
meillo@0 698 .LP
meillo@0 699 unix phil
meillo@0 700 .LP
meillo@0 701 case studies
meillo@0 702
meillo@0 703 .NH 2
meillo@0 704 Why people should choose
meillo@0 705 .LP
meillo@0 706 Make the right choice!
meillo@0 707
meillo@0 708 .nr PI .5i
meillo@0 709 .rm ]<
meillo@0 710 .de ]<
meillo@0 711 .LP
meillo@0 712 .de FP
meillo@0 713 .IP \\\\$1.
meillo@0 714 \\..
meillo@0 715 .rm FS FE
meillo@0 716 ..
meillo@0 717 .SH
meillo@0 718 References
meillo@0 719 .[
meillo@0 720 $LIST$
meillo@0 721 .]
meillo@0 722 .wh -1p