Re: [AD] April First

[ Thread Index | Date Index | More lists.liballeg.org/allegro-developers Archives ]


On Mon, 2009-02-23 at 17:15 -0800, Evert Glebbeek wrote:
> On 23 feb 2009, at 16:52, Peter Wang wrote:
> > A feature freeze is possible, but the deadline is too short to  
> > finalise
> > the API.
> 
> Yeah, it is coming very soon. We're close, I think, but the last few  
> bits always take the most time.
> 
> > That said, I have been thinking about a similar plan for later *this  
> > year*.
> > Instead of reaching for an unattainably perfect "5.0" release, we'd  
> > aim for
> > a decent 5.0 release, with an intent to release 5.2 relatively soon  
> > after;
> > the point of which is to nail down the API so people can start to  
> > use it
> > without worrying about getting the rug pulled from under their feet.
> > They may, however, hit bugs (unavoidable), implementation  
> > limitations, unclear
> > documentation, etc.  Of course, we should try to improve those  
> > aspects also,
> > but the focus would need to switch to critically reviewing the API  
> > that we
> > already have.
> 
> Oh, I agree, in general terms. "5.0" doesn't have to be perfect, but  
> it needs to be ok, without too many glaring bugs and the API has to be  
> reasonably complete and polished so people can really use it (we don't  
> want to release a KDE 4.0). However, real bugfixes and documentation  
> updates/clarifications go into 5.0.1, which I would see coming out a  
> month or two after 5.0.0 and then push for 5.2 within, say, a year or  
> half a year?

I also think this is the way to go. A5 is basically a new library
written from scratch, so it makes sense that we will find many and
sometimes quite glaring mistakes. So yes, the time between 5.0 and 5.2
could be significantly shorter than the normal time between two major
releasse (like e.g. 5.2 and 5.4). And also updating from 5.0 to 5.2
would probably require more work than from 5.2 to 5.4. But as we would
keeping making new 5.0.x releases through the lifetime of the 5.1
development branch, it would still be a good choice to start making
games with.

> 
> Something to keep in mind, now that I think of it, for when 5.0 is out  
> and we have extra time on our hands (hah!): many people are doing  
> development for portable devices (the iPhone is what I was mainly  
> thinking of here) and we should probably try to make it easy to get  
> Allegro programs ported there. In the particular case of the iPhone,  
> this might be relatively easy, since it runs a stripped-down version  
> of OS X. How hard it would be to do this depends on how stripped down  
> it really is.
> Anyway, something to keep in mind.

SDL got their iPhone port with a google summer of code project. So if
anyone has the extra time you speak of on their hands, could also use it
to fill out the google application forms and then select and mentor a
student to implement it. Whether that is more or less efficient than
using the same time to implement it yourself is another question of
course :)

> 
> Which reminds me of something else I forgot about: what happened to  
> the plan of having support for multiple mice?
> 

How would the API for it look like?

-- 
Elias Pschernig <elias@xxxxxxxxxx>





Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/