[osflash] Anybody interest in developing a new compiler?
robin at debreuil.com
Thu Sep 13 05:58:12 PDT 2007
>From all tests I made so far, declaring all functions with numbers
>instead of with names should speed up things.
Yes it does for sure, I've done it, and measured it, and it is many times
> Earlier you mentioned an optimization technique for method
> overloading. I don't think doing something like...
> myMethod_1( id:String );
That will still work if you store everything in tables (arrays). You can
also store the metadata separately or within the app (names, mangled names
etc), and then enable reflection. That can happen invisible to the user like
in the case of an imported 'dll' swf, or explicity (we implemented the
reflection framework from .net for that).
As for all the conversation about 'official' languages vs grassroots ones,
it reminds me of a quote from some dead linguist - "A language is a dialect
with an army and a navy". In this case though the 'army and navy' is
probably a good suite of tools, IDE and debugger ; ).
Personally I think the new JS specs aren't that great. It is trying to be
more scalable by adding a lot of Java type features, but it is still
dynamic, so it will never be scalable. AS2 suffers from the same problem, as
and got quite far imo (intellisense that was mostly complete and correct,
scriptdoc, debugging etc). But then to get to level 2, where you would trust
it to do a big refactor on your code (something you do all the time in
Java/C#), there is no way. The language allows too many crazy things to
happen to be able to trust automation to that level.
JS is a great language for quick and dirty projects, but like all dynamic
languages it scales poorly and is difficult to tool. So anyone working on
getting past that problem should be thanked imo. The issue with starting a
new language from scratch is really you also need to build out tools, IDE,
debugger etc. So it is a giant project in the end. If you implement an
existing language, you can piggyback in existing tools. I don't think that
means existing 'common' languages are better, but that is probably why they
As for choosing, it really isn't a big deal to learn a new familiar looking
language to code against the flash framework we already know, so why the
debate? It is probably a matter of a day or two before you would be up to
speed. For Haxe is it probably a matter of an hour or two : ). I really
don't see anyone having to choose one path or another.
More information about the osflash