Unnamed repository; edit this file to name it for gitweb.
git clone https://logand.com/git/picolisp.git/
Log | Files | Refs | README | LICENSE

faq.html (30365B)

      1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/1998/REC-html40-19980424/loose.dtd">
      2 <html lang="en">
      3 <head>
      4 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
      5 <title>PicoLisp FAQ</title>
      6 <link rel="stylesheet" href="doc.css" type="text/css">
      7 </head>
      8 <body>
      9 <a href="mailto:abu@software-lab.de">abu@software-lab.de</a>
     11 <p align=right>
     12 <i>Monk: "If I have nothing in my mind, what shall I do?"</i><br>
     13 <i>Joshu: "Throw it out."</i><br>
     14 <i>Monk: "But if there is nothing, how can I throw it out?"</i><br>
     15 <i>Joshu: "Well, then carry it out."</i><br>
     16 <i>(Zen koan)</i><br>
     18 <h1>PicoLisp Frequently Asked Questions</h1>
     20 <p align=right>(c) Software Lab. Alexander Burger
     22 <p><ul>
     23 <li><a href="#yet">Why did you write yet another Lisp?</a>
     24 <li><a href="#who">Who can use PicoLisp?</a>
     25 <li><a href="#advantages">What are the advantages over other Lisp systems?</a>
     26 <li><a href="#performance">How is the performance compared to other Lisp systems?</a>
     27 <li><a href="#interpreted">What means "interpreted"?</a>
     28 <li><a href="#compiler">Is there (or will be in the future) a compiler available?</a>
     29 <li><a href="#portable">Is it portable?</a>
     30 <li><a href="#webServer">Is PicoLisp a web server?</a>
     31 <li><a href="#lambda">I cannot find the LAMBDA keyword in PicoLisp</a>
     32 <li><a href="#dynamic">Why do you use dynamic variable binding?</a>
     33 <li><a href="#problems">Are there no problems caused by dynamic binding?</a>
     34 <li><a href="#closures">But with dynamic binding I cannot implement closures!</a>
     35 <li><a href="#macros">Do you have macros?</a>
     36 <li><a href="#strings">Why are there no strings?</a>
     37 <li><a href="#arrays">What about arrays?</a>
     38 <li><a href="#floats">How to do floating point arithmetics?</a>
     39 <li><a href="#bind">What happens when I locally bind a symbol which has a function definition?</a>
     40 <li><a href="#hardware">Would it make sense to build PicoLisp in hardware?</a>
     41 <li><a href="#segfault">I get a segfault if I ...</a>
     42 <li><a href="#ask">Where can I ask questions?</a>
     43 </ul>
     45 <p><hr>
     46 <h2><a name="yet">Why did you write yet another Lisp?</a></h2>
     48 <p>Because other Lisps are not the way I'd like them to be. They concentrate on
     49 efficient compilation, and lost the one-to-one relationship of language and
     50 virtual machine of an interpreted system, gave up power and flexibility, and
     51 impose unnecessary limitations on the freedom of the programmer. Other reasons
     52 are the case-insensitivity and complexity of current Lisp systems.
     55 <p><hr>
     56 <h2><a name="who">Who can use PicoLisp?</a></h2>
     58 <p>PicoLisp is for programmers who want to control their programming
     59 environment, at all levels, from the application domain down to the bare metal.
     60 Who want use a transparent and simple - yet universal - programming model, and
     61 want to know exactly what is going on. This is an aspect influenced by
     62 <code>Forth</code>.
     64 <p>It does <i>not</i> pretend to be easy to learn. There are already plenty of
     65 languages that do so. It is not for people who don't care what's under the hood,
     66 who just want to get their application running. They are better served with some
     67 standard, "safe" black-box, which may be easier to learn, and which allegedly
     68 better protects them from their own mistakes.
     71 <p><hr>
     72 <h2><a name="advantages">What are the advantages over other Lisp systems?</a></h2>
     74 <h3>Simplicity</h3>
     75 <p>PicoLisp is easy to understand and adapt. There is no compiler enforcing
     76 special rules, and the interpreter is simple and straightforward. There are only
     77 three data types: Numbers, symbols and lists ("LISP" means "List-, Integer- and
     78 Symbol Processing" after all ;-). The memory footprint is minimal, and the
     79 tarball size of the whole system is just a few hundred kilobytes.
     81 <h3>A Clear Model</h3>
     82 <p>Most other systems define the language, and leave it up to the implementation
     83 to follow the specifications. Therefore, language designers try to be as
     84 abstract and general as possible, leaving many questions and ambiguities to the
     85 users of the language.
     87 <p>PicoLisp does the opposite. Initially, only the single-cell data structure
     88 was defined, and then the structure of numbers, symbols and lists as they are
     89 composed of these cells. Everything else in the whole system follows from these
     90 axioms. This is documented in the chapter about the <a href="ref.html#vm">The
     91 PicoLisp Machine</a> in the reference manual.
     93 <h3>Orthogonality</h3>
     94 <p>There is only one symbolic data type, no distinction (confusion) between
     95 symbols, strings, variables, special variables and identifiers.
     97 <p>Most data-manipulation functions operate on the values of symbols as well as
     98 the CARs of cons pairs:
    100 <pre><code>
    101 : (let (N 7  L (7 7 7)) (inc 'N) (inc (cdr L)) (cons N L))
    102 -> (8 7 8 7)
    103 </code></pre>
    105 <p>There is only a single functional type, no "special forms". As there is no
    106 compiler, functions can be used instead of macros. No special "syntax"
    107 constructs are needed. This allows a completely orthogonal use of functions. For
    108 example, most other Lisps do not allow calls like
    110 <pre><code>
    111 : (mapcar if '(T NIL T NIL) '(1 2 3 4) '(5 6 7 8))
    112 -> (1 6 3 8)
    113 </code></pre>
    115 <p>PicoLisp has no such restrictions. It favors the principle of "Least
    116 Astonishment".
    118 <h3>Object System</h3>
    119 <p>The OOP system is very powerful, because it is fully dynamic, yet extremely
    120 simple:
    122 <p><ul>
    123 <li>In other systems you have to statically declare "slots". In PicoLisp,
    124 classes and objects are completely dynamic, they are created and extended at
    125 runtime. "Slots" don't even exist at creation time. They spring into existence
    126 purely dynamically. You can add any new property or any new method to any single
    127 object, at any time, regardless of its class.
    129 <li>The multiple inheritance is such that not only classes can have several
    130 superclasses, but each individual object can be of more than one class.
    132 <li>Prefix classes can surgically change the inheritance tree for any class or
    133 object. They behave like Mixins in this regard.
    135 <li>Fine-control of inheritance in methods with <code><a
    136 href="refS.html#super">super</a></code> and <code><a
    137 href="refE.html#extra">extra</a></code>.
    139 </ul>
    141 <h3>Pragmatism</h3>
    142 <p>PicoLisp has many practical features not found in other Lisp dialects. Among
    143 them are:
    145 <p><ul>
    146 <li>Auto-quoting of lists when the CAR is a number. Instead of <code>'(1 2
    147 3)</code> you can just write <code>(1 2 3)</code>. This is possible because a
    148 number never makes sense as a function name, and has to be checked at runtime
    149 anyway.
    151 <li>The <code><a href="refQ.html#quote">quote</a></code> function returns all
    152 unevaluated arguments, instead of just the first one. This is both faster
    153 (<code>quote</code> does not have to take the CAR of its argument list) and
    154 smaller (a single cell instead of two). For example, <code>'A</code> expands to
    155 <code>(quote . A)</code> and <code>'(A B C)</code> expands to <code>(quote A B
    156 C)</code>.
    158 <li>The symbol <code><a href="ref.html#atres">@</a></code> is automatically
    159 maintained as a local variable, and set implicitly in certain flow- and
    160 logic-functions. This makes it often unnecessary to allocate and assign local
    161 variables.
    163 <li><a href="tut.html#funio">Functional I/O</a> is more convenient than
    164 explicitly passing around file descriptors.
    166 <li>A well-defined <a href="ref.html#cmp">ordinal relationship</a> between
    167 arbitrary data types facilitates generalized comparing and sorting.
    169 <li>Uniform handling of <code>var</code> locations (i.e. values of symbols and
    170 CARs of cons pairs).
    172 <li>The universality and usefulness of symbol properties is enforced and
    173 extended with implicit and explicit bindings of the symbol <code><a
    174 href="refT.html#This">This</a></code> in combination with the access functions
    175 <code><a href="ref_.html#=:">=:</a></code>, <code><a
    176 href="ref_.html#:">:</a></code> and <code><a href="ref_.html#::">::</a></code>.
    178 <li>A very convenient list-building machinery, using the <code><a
    179 href="refL.html#link">link</a></code>, <code><a
    180 href="refY.html#yoke">yoke</a></code>, <code><a
    181 href="refC.html#chain">chain</a></code> and <code><a
    182 href="refM.html#made">made</a></code> functions in the <code><a
    183 href="refM.html#make">make</a></code> environment.
    185 <li>The syntax of often-used functions is kept non-verbose. For example, instead
    186 of <code>(let ((A 1) (B 2) C 3) ..)</code> you write <code>(let (A 1 B 2 C 3)
    187 ..)</code>, or just <code>(let A 1 ..)</code> if there is only a single
    188 variable.
    190 <li>The use of the hash (<code>#</code>) as a comment character is more adequate
    191 today, and allows a clean hash-bang (<code>#!</code>) syntax for stand-alone
    192 scripts.
    194 <li>The interpreter is <a href="ref.html#invoc">invoked</a> with a simple and
    195 flexible syntax, where command line arguments are either files to be interpreted
    196 or functions to be directly executed. With that, many tasks can be performed
    197 without writing a separate <a href="tut.html#script">script</a>.
    199 <li>A sophisticated system of interprocess communication, file locking and
    200 synchronization allows multi-user access to database applications.
    202 <li>A Prolog interpreter is tightly integrated into the language. Prolog
    203 clauses can call Lisp expressions and vice versa, and a self-adjusting
    204 depth-first search predicate <code>select</code> can be used in database
    205 queries.
    207 </ul>
    209 <h3>Persistent Symbols</h3>
    210 <p>Database objects ("external" symbols) are a primary data type in PicoLisp.
    211 They look like normal symbols to the programmer, but are managed in the database
    212 (fetched from, and stored to) automatically by the system. Symbol manipulation
    213 functions like <code>set</code>, <code>put</code> or <code>get</code>, the
    214 garbage collector, and other parts of the interpreter know about them.
    216 <h3>Application Server</h3>
    217 <p>It is a stand-alone system (it does not depend on external programs like
    218 Apache or MySQL) and it provides a "live" user interface on the client side,
    219 with an application server session for each connected client. The GUI layout and
    220 behavior are described with S-expressions, generated dynamically at runtime, and
    221 interact directly with the database structures.
    223 <h3>Localization</h3>
    224 <p>Internal exclusive and full use of UTF-8 encoding, and self-translating <a
    225 href="ref.html#transient-io">transient symbols</a> (strings), make it easy to
    226 write country- and language-independent applications.
    229 <p><hr>
    230 <h2><a name="performance">How is the performance compared to other Lisp systems?</a></h2>
    232 <p>Despite the fact that PicoLisp is an interpreted-only system, the performance
    233 is quite good. Typical Lisp programs operating on list data structures are
    234 executed in (interpreted) PicoLisp at about the same speed as in (compiled)
    235 CMUCL, and about two or three times faster than in CLisp or Scheme48. Programs
    236 with lots of numeric calculations, however, may be slower on a 32-bit system,
    237 due to PicoLisp's somewhat inefficient implementation of numbers. The 64-bit
    238 version improved on that.
    240 <p>But in practice, speed was never a problem, even with the first versions of
    241 PicoLisp in 1988 on a Mac II with a 12 MHz CPU. And certain things are cleaner
    242 and easier to do in plain <code>C</code> or <code>asm</code> anyway. It is very
    243 easy to write <code>C</code> functions in PicoLisp, either in the kernel, as
    244 shared object libraries, or even inline in the Lisp code.
    246 <p>PicoLisp is very space-effective. Other Lisp systems reserve heap space twice
    247 as much as needed, or use rather large internal structures to store cells and
    248 symbols. Each cell or minimal symbol in PicoLisp consists of only two pointers.
    249 No additional tags are stored, because they are implied in the pointer
    250 encodings. No gaps remain in the heap during allocation, as there are only
    251 objects of a single size. As a result, consing and garbage collection are very
    252 fast, and overall performance benefits from a better cache efficiency. Heap and
    253 stack grow automatically, and are limited only by hardware and operating system
    254 constraints.
    257 <p><hr>
    258 <h2><a name="interpreted">What means "interpreted"?</a></h2>
    260 <p>It means to directly execute Lisp data as program code. No transformation to
    261 another representation of the code (e.g. compilation), and no structural
    262 modifications of these data, takes place.
    264 <p>Lisp data are the "real" things, like numbers, symbols and lists, which can
    265 be directly handled by the system. They are <i>not</i> the textual
    266 representation of these structures (which is outside the Lisp realm and taken
    267 care of the <code><a href="refR.html#read">read</a></code>ing and <code><a
    268 href="refP.html#print">print</a></code>ing interfaces).
    270 <p>The following example builds a function and immediately calls it with two
    271 arguments:
    273 <pre><code>
    274 : ((list (list 'X 'Y) (list '* 'X 'Y)) 3 4)
    275 -> 12
    276 </code></pre>
    278 <p>Note that no time is wasted to build up a lexical environment. Variable
    279 bindings take place dynamically during interpretation.
    281 <p>A PicoLisp function is able to inspect or modify itself while it is running
    282 (though this is rarely done in application programming). The following function
    283 modifies itself by incrementing the '0' in its body:
    285 <pre><code>
    286 (de incMe ()
    287    (do 8
    288       (printsp 0)
    289       (inc (cdadr (cdadr incMe))) ) )
    291 : (incMe)
    292 0 1 2 3 4 5 6 7 -> 8
    293 : (incMe)
    294 8 9 10 11 12 13 14 15 -> 16
    295 </code></pre>
    297 <p>Only an interpreted Lisp can fully support such "Equivalence of Code and
    298 Data". If executable pieces of data are used frequently, like in PicoLisp's
    299 dynamically generated GUI, a fast interpreter is preferable over any compiler.
    302 <p><hr>
    303 <h2><a name="compiler">Is there (or will be in the future) a compiler available?</a></h2>
    305 <p>No. That would contradict the idea of PicoLisp's simple virtual machine
    306 structure. A compiler transforms it to another (physical) machine, with the
    307 result that many assumptions about the machine's behavior won't hold any more.
    308 Besides that, PicoLisp primitive functions evaluate their arguments
    309 independently and are not suited for being called from compiled code. Finally,
    310 the gain in execution speed would probably not be worth the effort. Typical
    311 PicoLisp applications often use single-pass code which is loaded, executed and
    312 thrown away; a process that would be considerably slowed down by compilation.
    315 <p><hr>
    316 <h2><a name="portable">Is it portable?</a></h2>
    318 <p>Yes and No. Though we wrote and tested PicoLisp originally only on Linux, it
    319 now also runs on FreeBSD, Mac OS X (Darwin), Cygwin/Win32, and probably other
    320 POSIX systems. The first versions were even fully portable between DOS, SCO-Unix
    321 and Macintosh systems. But today we have Linux. Linux itself is very portable,
    322 and you can get access to a Linux system almost everywhere. So why bother?
    324 <p>The GUI is completely platform independent (Browser), and in the times of
    325 Internet an application <u>server</u> does not really need to be portable.
    328 <p><hr>
    329 <h2><a name="webServer">Is PicoLisp a web server?</a></h2>
    331 <p>Not really, but it evolved a great deal into that direction.
    333 <p>Historically it was the other way round: We had a plain X11 GUI for our
    334 applications, and needed something platform independent. The solution was
    335 obvious: Browsers are installed virtually everywhere. So we developed a protocol
    336 which persuades a browser to function as a GUI front-end to our applications.
    337 This is much simpler than to develop a full-blown web server.
    340 <p><hr>
    341 <h2><a name="lambda">I cannot find the LAMBDA keyword in PicoLisp</a></h2>
    343 <p>Because it isn't there. The reason is that it is redundant; it is equivalent
    344 to the <code>quote</code> function in any aspect, because there's no distinction
    345 between code and data in PicoLisp, and <code>quote</code> returns the whole
    346 (unevaluated) argument list. If you insist on it, you can define your own
    347 <code>lambda</code>:
    349 <pre><code>
    350 : (def 'lambda quote)
    351 -> lambda
    352 : ((lambda (X Y) (+ X Y)) 3 4)
    353 -> 7
    354 : (mapcar (lambda (X) (+ 1 X)) '(1 2 3 4 5))
    355 -> (2 3 4 5 6)
    356 </code></pre>
    359 <p><hr>
    360 <h2><a name="dynamic">Why do you use dynamic variable binding?</a></h2>
    362 <p>Dynamic binding is very powerful, because there is only one single,
    363 dynamically changing environment active all the time. This makes it possible
    364 (e.g. for program snippets, interspersed with application data and/or passed
    365 over the network) to access the whole application context, freely, yet in a
    366 dynamically controlled manner. And (shallow) dynamic binding is the fastest
    367 method for a Lisp interpreter.
    369 <p>Lexical binding is more limited by definition, because each environment is
    370 deliberately restricted to the visible (textual) static scope within its
    371 establishing form. Therefore, most Lisps with lexical binding introduce "special
    372 variables" to support dynamic binding as well, and constructs like
    373 <code>labels</code> to extend the scope of variables beyond a single function.
    375 <p>In PicoLisp, function definitions are normal symbol values. They can be
    376 dynamically rebound like other variables. As a useful real-world example, take
    377 this little gem:
    379 <pre><code>
    380 (de recur recurse
    381    (run (cdr recurse)) )
    382 </code></pre>
    384 <p>It implements anonymous recursion, by defining <code>recur</code> statically
    385 and <code>recurse</code> dynamically. Usually it is very cumbersome to think up
    386 a name for a function (like the following one) which is used only in a single
    387 place. But with <code>recur</code> and <code>recurse</code> you can simply
    388 write:
    390 <pre><code>
    391 : (mapcar
    392    '((N)
    393       (recur (N)
    394          (if (=0 N)
    395             1
    396             (* N (recurse (- N 1))) ) ) )
    397    (1 2 3 4 5 6 7 8) )
    398 -> (1 2 6 24 120 720 5040 40320)
    399 </code></pre>
    401 <p>Needless to say, the call to <code>recurse</code> does not have to reside in
    402 the same function as the corresponding <code>recur</code>. Can you implement
    403 anonymous recursion so elegantly with lexical binding?
    406 <p><hr>
    407 <h2><a name="problems">Are there no problems caused by dynamic binding?</a></h2>
    409 <p>You mean the <i>funarg</i> problem, or problems that arise when a variable
    410 might be bound to <i>itself</i>? For that reason we have a convention in
    411 PicoLisp to use <a href="ref.html#transient-io">transient symbols</a> (instead
    412 of internal symbols)
    414 <ol>
    416 <li>for all parameters and locals, when functional arguments or executable lists
    417 are passed through the current dynamic bindings
    419 <li>for a parameter or local, when that symbol might possibly be (directly or
    420 indirectly) bound to itself, and the bound symbol's value is accessed in the
    421 dynamic context
    423 </ol>
    425 <p>This is a form of lexical <i>scoping</i> - though we still have dynamic
    426 <i>binding</i> - of symbols, similar to the <code>static</code> keyword in
    427 <code>C</code>.
    429 <p>In fact, these problems are a real threat, and may lead to mysterious bugs
    430 (other Lisps have similar problems, e.g. with symbol capture in macros). They
    431 can be avoided, however, when the above conventions are observed. As an example,
    432 consider a function which doubles the value in a variable:
    434 <pre><code>
    435 (de double (Var)
    436    (set Var (* 2 (val Var))) )
    437 </code></pre>
    439 <p>This works fine, as long as we call it as <code>(double 'X)</code>, but will
    440 break if we call it as <code>(double 'Var)</code>. Therefore, the correct
    441 implementation of <code>double</code> should be:
    443 <pre><code>
    444 (de double ("Var")
    445    (set "Var" (* 2 (val "Var"))) )
    446 </code></pre>
    448 <p>If <code>double</code> is defined that way in a separate source file, and/or
    449 isolated via the <code><a href="ref_.html#====">====</a></code> function, then
    450 the symbol <code><u>Var</u></code> is locked into a private lexical context
    451 and cannot conflict with other symbols.
    453 <p>Admittedly, there are two disadvantages with this solution:
    455 <ol>
    457 <li>The rules for when to use transient symbols are a bit complicated. Though it
    458 is safe to use them even when not necessary, it will take more space then and be
    459 more difficult to debug.
    461 <li>The string-like syntax of transient symbols as variables may look strange to
    462 alumni of other languages.
    464 </ol>
    466 Fortunately, these pitfalls do not occur so very often, and seem more likely in
    467 utilities than in production code, so that they can be easily encapsulated.
    470 <p><hr>
    471 <h2><a name="closures">But with dynamic binding I cannot implement closures!</a></h2>
    473 <p>This is not true. Closures are a matter of scope, not of binding.
    475 <p>For a closure it is necessary to build and maintain a separate environment.
    476 In a system with lexical bindings, this has to be done at <i>each</i> function
    477 call, and for compiled code it is the most efficient strategy anyway, because it
    478 is done once by the compiler, and can then be accessed as stack frames at
    479 runtime.
    481 <p>For an interpreter, however, this is quite an overhead. So it should not be
    482 done automatically at each and every function invocation, but only if needed.
    484 <p>You have several options in PicoLisp. For simple cases, you can take
    485 advantage of the static scope of <a href="ref.html#transient-io">transient
    486 symbols</a>. For the general case, PicoLisp has built-in functions like <code><a
    487 href="refB.html#bind">bind</a></code> or <code><a
    488 href="refJ.html#job">job</a></code>, which dynamically manage statically scoped
    489 environments.
    491 <p>Environments are first-class objects in PicoLisp, more flexible than
    492 hard-coded closures, because they can be created and manipulated independently
    493 from the code.
    495 <p>As an example, consider a currying function:
    497 <pre><code>
    498 (de curry Args
    499    (list (car Args)
    500       (list 'list
    501          (lit (cadr Args))
    502          (list 'cons ''job
    503             (list 'cons
    504                (list 'lit (list 'env (lit (car Args))))
    505                (lit (cddr Args)) ) ) ) ) )
    506 </code></pre>
    508 <p>When called, it returns a function-building function which may be applied to
    509 some argument:
    511 <pre><code>
    512 : ((curry (X) (N) (* X N)) 3)
    513 -> ((N) (job '((X . 3)) (* X N)))
    514 </code></pre>
    516 <p>or used as:
    518 <pre><code>
    519 : (((curry (X) (N) (* X N)) 3) 4)
    520 -> 12
    521 </code></pre>
    523 <p>In other cases, you are free to choose a shorter and faster solution. If (as
    524 in the example above) the curried argument is known to be immutable:
    526 <pre><code>
    527 (de curry Args
    528    (list
    529       (cadr Args)
    530       (list 'fill
    531          (lit (cons (car Args) (cddr Args)))
    532          (lit (cadr Args)) ) ) )
    533 </code></pre>
    535 <p>Then the function built above will just be:
    537 <pre><code>
    538 : ((curry (X) (N) (* X N)) 3)
    539 -> ((X) (* X 3))
    540 </code></pre>
    542 <p>In that case, the "environment build-up" is reduced by a simple (lexical)
    543 constant substitution with zero runtime overhead.
    545 <p>Note that the actual <code><a href="refC.html#curry">curry</a></code>
    546 function is simpler and more pragmatic. It combines both strategies (to use
    547 <code>job</code>, or to substitute), deciding at runtime what kind of function
    548 to build.
    551 <p><hr>
    552 <h2><a name="macros">Do you have macros?</a></h2>
    554 <p>Yes, there is a macro mechanism in PicoLisp, to build and immediately execute
    555 a list of expressions. But it is seldom used. Macros are a kludge. Most things
    556 where you need macros in other Lisps are directly expressible as functions in
    557 PicoLisp, which (as opposed to macros) can be applied, passed around, and
    558 debugged.
    560 <p>For example, Common Lisp's <code>DO*</code> macro, written as a function:
    562 <pre><code>
    563 (de do* "Args"
    564    (bind (mapcar car (car "Args"))
    565       (for "A" (car "Args")
    566          (set (car "A") (eval (cadr "A"))) )
    567       (until (eval (caadr "Args"))
    568          (run (cddr "Args"))
    569          (for "A" (car "Args")
    570             (and (cddr "A") (set (car "A") (run @))) ) )
    571       (run (cdadr "Args")) ) )
    572 </code></pre>
    575 <p><hr>
    576 <h2><a name="strings">Why are there no strings?</a></h2>
    578 <p>Because PicoLisp has something better: <a
    579 href="ref.html#transient-io">Transient symbols</a>. They look and behave like
    580 strings in any respect, but are nevertheless true symbols, with a value and a
    581 property list.
    583 <p>This leads to interesting opportunities. The value, for example, can point to
    584 other data that represent the string's translation. This is used extensively for
    585 localization. When a program calls
    587 <pre><code>
    588    (prinl "Good morning!")
    589 </code></pre>
    591 <p>then changing the value of the symbol <code>"Good morning!"</code> to its
    592 translation will change the program's output at runtime.
    594 <p>Transient symbols are also quite memory-conservative. As they are stored in
    595 normal heap cells, no additional overhead for memory management is induced. The
    596 cell holds the symbol's value in its CDR, and the tail in its CAR. If the string
    597 is not longer than 7 bytes, it fits (on the 64-bit version) completely into the
    598 tail, and a single cell suffices. Up to 15 bytes take up two cells, 23 bytes
    599 three etc., so that long strings are not very efficient (needing twice the
    600 memory on the average), but this disadvantage is made up by simplicity and
    601 uniformity. And lots of extremely long strings are not the common case, as they
    602 are split up anyway during processing, and stored as plain byte sequences in
    603 external files and databases.
    605 <p>Because transient symbols are temporarily interned (while <code><a
    606 href="refL.html#load">load</a></code>ing the current source file), they are
    607 shared within the same source and occupy that space only once, even if they
    608 occur multiple times within the same file.
    611 <p><hr>
    612 <h2><a name="arrays">What about arrays?</a></h2>
    614 <p>PicoLisp has no array or vector data type. Instead, lists must be used for
    615 any type of sequentially arranged data.
    617 <p>We believe that arrays are usually overrated. Textbook wisdom tells that they
    618 have a constant access time O(1) when the index is known. Many other operations
    619 like splits or insertions are rather expensive. Access with a known (numeric)
    620 index is not really typical for Lisp, and even then the advantage of an array is
    621 significant only if it is relatively long. Holding lots of data in long arrays,
    622 however, smells quite like a program design error, and we suspect that often
    623 more structured representations like trees or interconnected objects would be
    624 better.
    626 <p>In practice, most arrays are rather short, or the program can be designed in
    627 such a way that long arrays (or at least an indexed access) are avoided.
    629 <p>Using lists, on the other hand, has advantages. We have so many concerted
    630 functions that uniformly operate on lists. There is no separate data type that
    631 has to be handled by the interpreter, garbage collector, I/O, database and so
    632 on. Lists can be made circular. And lists don't cause memory fragmentation.
    635 <p><hr>
    636 <h2><a name="floats">How to do floating point arithmetics?</a></h2>
    638 <p>PicoLisp does not support real floating point numbers. You can do all kinds
    639 of floating point calculations by calling existing library functions via
    640 <code><a href="refN.html#native">native</a></code>, inline-C code, and/or by
    641 loading the "@lib/math.l" library.
    643 <p>But PicoLisp has something even (arguably) better: Scaled <a
    644 href="ref.html#num-io">fixpoint numbers</a>, with unlimited precision.
    646 <p>The reasons for this design decision are manifold. Floating point numbers
    647 smack of imperfection, they don't give "exact" results, have limited precision
    648 and range, and require an extra data type. It is hard to understand what really
    649 goes on (How many digits of precision do we have today? Are perhaps 10-byte
    650 floats used for intermediate results? How does rounding behave?).
    652 <p>For fixpoint support, the system must handle just integer arithmetics, I/O
    653 and string conversions. The rest is under programmer's control and
    654 responsibility (the essence of PicoLisp).
    656 <p>Carefully scaled fixpoint calculations can do anything floating points can
    657 do.
    660 <p><hr>
    661 <h2><a name="bind">What happens when I locally bind a symbol which has a function definition?</a></h2>
    663 <p>That's not a good idea. The next time that function gets executed within the
    664 dynamic context the system may crash. Therefore we have a convention to use an
    665 upper case first letter for locally bound symbols:
    667 <pre><code>
    668 (de findCar (Car List)
    669    (when (member Car (cdr List))
    670       (list Car (car List)) ) )
    671 </code></pre>
    673 ;-)
    676 <p><hr>
    677 <h2><a name="hardware">Would it make sense to build PicoLisp in hardware?</a></h2>
    679 <p>At least it should be interesting. It would be a machine executing list
    680 (tree) structures instead of linear instruction sequences. "Instruction
    681 prefetch" would look down the CAR- and CDR-chains, and perhaps need only a
    682 single cache for both data and instructions.
    684 <p>Primitive functions like <code>set</code>, <code>val</code>, <code>if</code>
    685 and <code>while</code>, which are written in <code>C</code> or assembly language
    686 now, would be implemented in microcode. Plus a few I/O functions for hardware
    687 access. <code>EVAL</code> itself would be a microcode subroutine.
    689 <p>Only a single heap and a single stack is needed. They grow towards each
    690 other, and cause garbage collection if they get too close. Heap compaction is
    691 trivial due to the single cell size.
    693 <p>There would be no assembly-language. The lowest level (above the hardware and
    694 microcode levels) are s-expressions: The machine language is <i>Lisp</i>.
    697 <p><hr>
    698 <h2><a name="segfault">I get a segfault if I ...</a></h2>
    700 <p>It is easy to produce a segfault in PicoLisp. Just set a symbol to a value
    701 which is not a function, and call it:
    703 <pre><code>
    704 : (setq foo 1)
    705 -> 1
    706 : (foo)
    707 Segmentation fault
    708 </code></pre>
    710 There is another <a href="ref.html#codePointer">example</a> in the <a
    711 href="ref.html#ev">Evaluation</a> section of the reference manual.
    713 <p>PicoLisp is a pragmatic language. It doesn't check at runtime for all
    714 possible error conditions which won't occur during normal usage. Such errors are
    715 usually detected quickly at the first test run, and checking for them after that
    716 would just produce runtime overhead.
    718 <p>Catching the segmentation violation and bus fault signals is also not a good
    719 idea, because the Lisp heap is most probably be damaged afterwards, possibly
    720 creating further havoc if execution continues.
    722 <p>It is recommended to inspect the code periodically with <code><a
    723 href="refL.html#lint">lint</a></code>. It will detect many potential errors.
    724 And, most of these errors are avoided by following the PicoLisp <a
    725 href="ref.html#conv">naming conventions</a>.
    728 <p><hr>
    729 <h2><a name="ask">Where can I ask questions?</a></h2>
    731 <p>The best place is the <a
    732 href="mailto:picolisp@software-lab.de?subject=Subscribe">PicoLisp Mailing
    733 List</a> (see also <a
    734 href="http://www.mail-archive.com/picolisp@software-lab.de/">The Mail
    735 Archive</a> and <a
    736 href="http://dir.gmane.org/gmane.lisp.picolisp.general">Gmane.org</a>), or the
    737 IRC <a href="irc://irc.freenode.net/picolisp">#picolisp</a> channel on
    738 FreeNode.net.
    740 </body>
    741 </html>