On 24/6/99 3:09 am spierings <[EMAIL PROTECTED]> wrote:
>> > Has any of the decisions made on this group been put into formal
>> > documentation. If not, it really needs to be. If I want to write a
>> > compiler to conform to the xTalk standard, how do I go about it?
>>
>> The closest thing to a standard in this area is HyperCard 2.4.1.
>> Though HyperTalk only supports a subset of what is available in the
>> other xTalk products, what it does have is clearly the core of what
>> would be in a standard. Unfortunately, that implementation, just like
>> all the others, has its weaknesses and ambiguities and since it's not
>> even being developed anymore the odds of being able to use it as a the
>> basis of a real standard are very small (no language implementation
>> has ever made the transition to being a standard unscathed ;-)
>> Regards,
>> Scott
>
>I thought this group was for defining a standard for xTalk so all
>compilers could compile a script if it conformed to the standard. This
>seems a really futile exercise if this discussion is not made a formal
>language specification.
>
>Hmmm... so what is this list for exactly again?
When the list was originally set up, standardisation was the intention.
However, since then the number of xTalk vendors has dropped. OMO is
dead, HC is showing few signs of life, the SuperCard authors have
expressed little interest in this list, Hypersense appears to be ticking
over, which leaves MetaCard. So it became rather pointless to write
languge specifications. But that position *may* have changed given the
Opencard initiative. Though I'm not sure if a specification only in use
by two vendors (or three if we're lucky) is necessarily worth the effort
of writing it.
>If we want to standardize xTalk why don't we do it properly. Does anybody
>have experience in writing language specifications?
>
>> (no language implementation has ever made the transition to being a
>standard unscathed ;-)
>
>There would only be required elements to call your language xTalk. You
>could add other stuff and still conform. Maybe a standard library of
>internal functions like :put, cos, sin. And flow control if e then; end
>if. Would be all you would need to conform.
Unfortunately, we don't currently have the time for a project like this.
Too busy writing software ;-) Its a sizeable task, and requires someone
with a fair bit of available time. Perhaps someone from the Opencard
group would like to do it?
Regards,
Kevin
>Andre
Kevin Miller <[EMAIL PROTECTED]> <http://www.xworlds.com/>
Cross Worlds Computing, MetaCard Distributors, Custom Development.
Tel: +44 (0)131 672 2909. Fax: +44 (0)1639 830 707.
