Re: Exploding feature set...time to apply the brakes?

"Mukhsein Johari" <[email protected]>
> I'd like to ask if we really want all the proposed
> features by 1.0? Could we not defer some till 2.0
> maybe? Especially the larger or more esoteric features
> like 'decorators'. My concern is that with too many
> features (and keywords), it will take longer to get
> stable or even debug effectively. More features ==
> more features to debug.

Whether a good decision or not, last year I made the decision to design 1.0
features until now.  That is why it is the way it is.  We are now closing up
the process.  So in a sense you are getting your wish.  The list on the Wiki
is amazingly short.  You should have been around here a month ago.  :-)

> There are other things to do (modules?), right now
> it's still pretty much 'batteries not included' as far
> as day-to-day usage is concerned. We need a networking
> module, to be sure. And some xml-related stuff. The
> earlier we have batteries the earlier prothon is
> useful, the earlier we get more user feedback. It's
> tough to build modules when the core is so fluid.

The next phase is making it reliable and optimizing it.  Fixing the memory
leaks, garbage collection, going to a jit, etc.  Hopefully in parallel we
can start adding batteries as it gets more reliable and since the language
features will be settled.

> Aside:
> I like the way twisted implements networking and
> protocols. This is not the same as saying I like the
> twisted implementation. Whether or not we need async
> networking will depend on whether prothon has the same
> problems as python when it comes to threading and so
> on. The separation of transport and (application)
> protocol is a good idea though.

Our threading should be excellent, right?