[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
Evert Glebbeek wrote:
Would putting it as include/allegro/oldapi.h sound like a good idea? For
now, it would only contain the old rest. For the 4.2 release, it will
contain all the functions, with the prefix #defined away, as well as all
deprecated material.
I'm all for doing this for 4.2, but I don't think it should go into WIP 4.1.15
just for rest(). IMO, this is something to start doing after the upcoming WIP
release.
This will also set a clear goal for the 4.1.16 WIP: present a prefixed API in
preparation for the final 4.2 release.
I'm curious what others (Eric, Peter) think about this.
I think we're entering into an infinite loop.
1. Someone suggests prefixing the whole API.
2. Well, it's kinda pointless just to prefix the API. Since we're breaking
compatibility anyway, let's fix up the uglier parts of the API at the
same time.
3. Hey, we only get one shot at this, and it has to be 100% absolutely
completely perfect! Let's restart from scratch... Allegro 5! Hurrah!
4. Development of Allegro 4 stagnates. Patches and improvements to
Allegro 4 are delayed or forgotten. True believers urge "just wait until
Allegro 5 is done, it will be AWESOME". Allegro 5 itself never gets any
momentum. Months pass.
5. Some people finally get back to Allegro 4.
6. After not too long, go to 1.
If I was to go back in time, my advice would be:
1. Prefixing is overrated, get over it. Really.
2. Improve the API incrementally. Maintain compatibility wherever
reasonable.
3. If you want design an API from scratch, DON'T CALL IT "ALLEGRO" and
DON'T CLAIM TO BE A SUCCESSOR. Either it is a complete lie, or you risk
starving necessary resources from the original. Build up your own
community
(this is the hard part, not the coding). Don't be a parasite.
Ok, no one asked for my advice. Just promise me you guys will break out
of the
loop before step 3.
Peter