>Alain: What a drag. :( I was hoping that our
>standalones could automatically shed everything they
>don't need, in order to be as small as they can be.
Alain,
automatically? I doubt this would be effective, as e.g. a stack that uses
the "set script" command would cause problems when it comes to detecting
what commands are used and might lead to exclusion of other commands. OTOH
checking variable contents might lead to e.g. a FreeScript Reference stack
standalone to include too many commands etc...
>Anthony: We could do this for some platforms, such as
>Unix.
>
>Alain: The Mac is always the renegade, eh!
>Ironically, their 'different' way of doing things is a
>hindrance to you.
It could be done on the Mac, too, but it would involve platform-specific
code that deeply messes with the intrinsics of our application. We could
e.g. compile every command into its own code resource, which would allow
just copying over what we need etc. However, I thought cross-platform code
was our first goal, and it'll definitely hamper portability if we add such
complicated platform-specific stuff, especially as it can't be factored out
as easily; it has to be involved with every command!
>Perhaps we should consider
>developing FreeCard for Unix/Linux, THEN port it to
>Mac. It would make our development efforts more
>focused. We could release the Linux-only version
>quicker and more efficiently than we can release a
>multi-platform version that includes Linux. The
>release would generate interest, attract new users and
>developers. It might even attract investors who would
>view the released Linux-only version as a
>proof-of-concept.
Fine. We'd just need some Linux programmers that would do this. I posted a
request to the two main Linux and Windows programming newsgroups and got no
response at all. I guess they want something they can see before they'll
believe and start contributing and porting.
>Anthony: We'd have to make very liberal use of the
>linker, though.
>
>Alain: Is that a problem?
1) It would bind FreeCard to an external program for this purpose that
isn't under our control and we'd have to make certain it's present
2) It would be a platform-specific aspect
3) it wouldn't solve all our problems because we can't just 'leave out'
some parts without causing dangerous holes.
There are ways to solve these problems, I guess, but they would all have
to be done very carefully or we'll have problems porting or would decrease
stability.
Cheers,
-- M. Uli Kusterer
------------------------------------------------------------
http://www.weblayout.com/witness
'The Witnesses of TeachText are everywhere...'
