Re: any emulation docs ?

[pausch@saaf.se (Paul Schlyter)]

In article <fakpdh$2iai$1@ihnp4.ucsd.edu>,
Jake Wildstrom <dwildstr@zeno.ucsd.edu> wrote:
>The Prophet Paul Schlyter, known to the wise as pausch@saaf.se, opened the Book of Words, and read unto the people:
>>In article <1187887854.475779.187080@m37g2000prh.googlegroups.com>,
>> <zwsdotcom@gmail.com> wrote:
>>
>>> No system ever built, past, present or future, has *full*
>>> documentation.
>>> 
>>> Most of the challenge in writing an emulator is finding and emulating
>>> undocumented behaviors.
>>
>>What you say apply only to real systems, i.e. systems already
>>implemented in hardware.  But kestas wanted to write an emulator of
>>"any system", i.e. not necessarily any existing system.  That makes him
>>free to invent any system he wants, produce documentation as complete
>>as he wants for his imaginary system, and then implement an emulator
>>for it.  When he's implemented an emulkator for it, his system isn't
>>quite as imaginary anymore since it has an implementation.
>
>Well, even virtual machines have undiscovered warts, if they've
>developed over time. Take, for instance, item 8.8.3.2.2.1 of the
>Z-machine standards:
>
>  If the interpreter number is 6 (for MSDOS) and the story file
>  is 'Zork Zero' release 393.890714, but in no other case, the
>  interpreter must do the following instead [of the ordinary newline
>  routine]: (1) move to the new line, (2) put the cursor at the
>  current left margin, (3) call the interrupt routine (if it's time to
>  do so).
>
>I daresay Java has undocumented behaviors at this point. Or at least
>particular implementations of the Java VM do.

If the JVM specification says the behaviour is undocumented at such
a point, then that undocumented behaviour is documented.... <g>



-- 
----------------------------------------------------------------
Paul Schlyter,  Grev Turegatan 40,  SE-114 38 Stockholm,  SWEDEN
e-mail:  pausch at stockholm dot bostream dot se
WWW:     http://stjarnhimlen.se/