search

HaxeFoundation / haxe.org

The haxe.org website
https://haxe.org
81 stars 96 forks source link

Target tier list #407

Closed Simn closed 4 years ago

Simn commented 5 years ago

I would like to document a target tier list which indicates how much focus we put on individual targets. My suggestion:

Tier 1

Actively maintained by the core compiler team with a strong focus from the Haxe Foundation

Tier 2

Maintained mostly by individuals in their free time

Tier 3

Kept up-to-date, but not much active development

Tier 4

Essentially deprecated

Aurel300 commented 5 years ago

As a new column here?

ncannasse commented 5 years ago

I think it's a bit misleading and does not directly translate to the quality of the target itself. For instance the Neko target, while almost deprecated, works very well with almost no known bugs (apart from missing Unicode support). Whereas there's much more known issues in C# target for example.

So I think a better way would be to have a single line of "current status" for each target, which does not exactly translate into tiers, but gives a more precise information regarding the target status and future evolution.

ncannasse commented 5 years ago

For instance, for my own targets, that would be:

HashLink : stable, ongoing development Neko : stable, lacks Unicode support, will be deprecated in medium term in favor of HashLink and Eval

Simn commented 5 years ago

My motivation here is to steer people towards using higher tier targets given a choice, which we can achieve nicely this way.

Not sure if we really have to go into much more detail here. I like the idea of being able to say "Look, this target is only tier 3 support, so fixes may take a long time". I find that much more tangible than some notion of "stable".

ibilon commented 5 years ago

Stable and bug-free is dependent on the amount of use the target has, less used ones will have less bug report.

And we can assume that all the targets have a least one bug, and that one bug will bother at least one user. A good information is: will it be fixed and how soon?

Like the c# and java target have a whole class of potential or reported bugs with type parameters and casting, but nobody really want to touch that can of worms.

Tier of support can inform target choice imo.

nanjizal commented 5 years ago

C++ needs to be a Tier 1 priority ?