Sam made reference to the CLS last week, and it's a point well-taken. We need a list of standard activities that parrot provides so that people writing compilers can use the provided facilities and people implementing data types can make sure they do the right things to work with the standards.
Before I go rattling this off and starting a PDD that'll likely languish a bit, does anyone want to step up and volunteer to put together a skeleton PDD with the current facilities, at which point we can start moving forward to fill in any holes and getting things detailed enough that it can be reasonably used?
To start it'd include the vtable and MMD functions, the ops that call 'em, and the methods we look for by default. -- Dan
--------------------------------------it's like this------------------- Dan Sugalski even samurai d...@sidhe.org have teddy bears and even teddy bears get drunk
Dan Sugalski wrote: > Sam made reference to the CLS last week, and it's a point well-taken. We > need a list of standard activities that parrot provides so that people > writing compilers can use the provided facilities and people > implementing data types can make sure they do the right things to work > with the standards.
> Before I go rattling this off and starting a PDD that'll likely languish > a bit, does anyone want to step up and volunteer to put together a > skeleton PDD with the current facilities, at which point we can start > moving forward to fill in any holes and getting things detailed enough > that it can be reasonably used?
> To start it'd include the vtable and MMD functions, the ops that call > 'em, and the methods we look for by default.
My goal at the moment is to get up to speed to the point where I *could* write such a document. I'm actively working on Pirate, have been in concact with the authors of Pint, and also plan to look into a Parrot binding to libxml2 (which would explain my interest in utf-8).
Note: while I was in no way the author of the CLA, I was the conveener (EMCA-speak for chair) of the working group which produced it, so I am familiar with both its content and how it was produced.
With that in mind, it is my feeling that such a document is premature. Unless somebody is much further along than I am aware of, we don't know where all the pitfalls are just yet; let alone knowing which of the pitfalls we can pave over, and which we simply need to post a big bold lighthouse at.
If you can wait a few weeks, I'm willing to take a stab at it, but just not yet. If somebody else takes the lead, I will certainly contribute.
On Tue, 09 Nov 2004 12:09:26 -0500, Sam Ruby <ru...@intertwingly.net> wrote: > My goal at the moment is to get up to speed to the point where I *could* > write such a document.
Indeed. Being able to eventually write docs is much of the reason I've been working on a Forth implementation. I think Sam has eclipsed me though in terms of understanding/achievement.
Once I'm able to get Forth working (close to 100%), I plan to document what I've learned from experimentation, the list, and IRC. There's certainly a lot to be done.
>>Sam made reference to the CLS last week, and it's a point >>well-taken. We need a list of standard activities that parrot >>provides so that people writing compilers can use the provided >>facilities and people implementing data types can make sure they do >>the right things to work with the standards.
>>Before I go rattling this off and starting a PDD that'll likely >>languish a bit, does anyone want to step up and volunteer to put >>together a skeleton PDD with the current facilities, at which point >>we can start moving forward to fill in any holes and getting things >>detailed enough that it can be reasonably used?
>>To start it'd include the vtable and MMD functions, the ops that >>call 'em, and the methods we look for by default.
>With that in mind, it is my feeling that such a document is >premature. Unless somebody is much further along than I am aware of, >we don't know where all the pitfalls are just yet; let alone knowing >which of the pitfalls we can pave over, and which we simply need to >post a big bold lighthouse at.
I can certainly wait. I'm not looking for a final document now, though, more a detailing of what we currently do. There are a lot of core facilites that we really want people writing compilers to use, but until they're detailed it's mostly luck if they are. (Well, and laziness -- we can be reasonably sure compilers will emit a mul op to do multiplication, but it's good to be explicit) -- Dan
--------------------------------------it's like this------------------- Dan Sugalski even samurai d...@sidhe.org have teddy bears and even teddy bears get drunk
|
