<<Stable enough for people to write/fix plugins without expecting any earth shattering changes.>>


Well frankly the changes weren't even that huge between DM and ToD. All we're talking about here are a few different function prototypes and locations, and a few things you have to do slightly differently. People that used ImpFilter had to find the functions somewhere else. That really isn't THAT big of a deal.

 

-----signature-----

<< Well frankly the changes weren't even that huge between DM and ToD. >>


Depends. HUD's were totally broken. Imp Filter went away (big one for lots of people), World Filter got changed.

 

-----signature-----

*gets on soapbox*


Haz, bite me.


Let me direct you to Decal's own site and the public announcement of Alpha 4:


http://forums.acdev.org/phpBB2/viewtopic.php?t=3627


In particular, let me direct your attention to this part:


"There are still many things that we've not finished, and much work still to do, and I can guarantee this won't be the last time we have breaking changes, since we still haven't completed our rework of the rendering system."


Does that sound stable to you? Does the fact that its still ALPHA sound stable? Obviously if the Voice of Decal comes out and publically declares that the so-called 'stable' version is NOT the last of the breaking changes, then its NOT stable.


Oh wait, we have V5 out now I guess...but lets see what thread we have here...oh yes, the one about breaking compile in many languages, and discussing 5A, 5B, and 5C. For reference:


(URL available if needed, but not sure its public forum so I'll skip it)


How exactly is that 'stable'?


Oh wait, my favorite quote, from YOU nonetheless...(URL available if needed, but not sure its public forum so I'll skip it)


"ou were having issues moving your development base to Alpha 5, but I suggest you do so. This is a FAR more stable version of Decal than Alpha 3 or Alpha 4, and provided none of you developers get crashing bugs after recompiling your plugins, we're going to post it."


*off soapbox*


Now, let me pull back just a bit from playing overly-zealous-pedantic-devils-advocate and just say that I really don't mind the changes as much as I mind someone telling me what I can and cannot perceive and communicate.


p.s. yes, I understand you are saying the API has been stable since Alpha 4, but that is not the sole facet of stability and you know it.

 

-----signature-----

You don't like being told what to communicate? Works both ways, I'm tired of people maligning Decal as a massively unstable piece of code ready to explode at any moment. It's not.

 

-----signature-----

I don't think the perception of Decal is as bad as you think it is. At least *I* don't take it (or intend it) the way you appear to be taking it. I still consider (*NDA*) unstable and its been through several betas as well as three alphas. If you want Decal to be generally percieved as stable, you are going to have to go more than one release without making developers change development environment versions, breaking all plugins, or putting out an A, B, and C version of the same release. A decent plugin framework should exist(to show people how to 'properly' do things the new way, e.g. adapter) and the tools should be updated to work with the new way of doing things. Decal used to have all this, in multiple languages. Its starting to have them available again, but these are very very recent additions and its going to take people some time to reevaluate the state of the Decal programming environment and call it stable.


I think what you call 'stability' I refer to as 'maturity'. The blueprint may be set and the foundation may be in place, but I dont consider the house 'stable'/'mature' until the drywall is up and the carpets are down.

 

-----signature-----

Well, would you rather we had several release candidates or a buggy final Alpha 5? It's not like end users will be seeing any of that, they will get Alpha 5 and that will be the end of it.


And when have we EVER required authors change development environments? VB6 still works, and we even support it as a development target. And I don't think recompilation between Alpha versions is overly onerous.

 

-----signature-----

" And when have we EVER required authors change development environments?"


Adapter adopters(catchy phrase, that!) have to match specific versions. I can't just use VB6(correct me if I'm wrong but wasnt official support for VB6 dropped?), or even VS2003. Every time you change revs, I have to change mine. Obviously, I don't *have* to use adapter, but I don't *have* to use anything at all. I *could* write (Lord help me) assembly COM calls to the interops, and I *could* just write all my code in notepad and ship it to someone who has the adapter to build, but lets get real. The general attitude is that the adapter is the preferred method and I take that at face value when considering which approach to use in development. I really do try to listen to you guys and when you say 'this is a better way to do X or Y', I believe you. When faced with a VB6 ImpFilter-based plugin to update, I don't really have any practical alternative other than to change environments, especially if I want to get support/frameworks/etc.

 

-----signature-----

No, VB6 is still supported. If VB6 wasn't supported, it wouldn't be possible to do much in a plugin with it. We've had to change things numerous times so Decal would even work with VB6. VS2K3 .NET (1.0/1.1) still works fine, though you have to use full COM interop rather than Adapter (since Adapter is a .NET 2.0 construct). C# and VB.NET Express editions are free, so we have no real qualms with requiring it for Adapter.


Adapter is being pushed because:

1) It's much easier to write a plugin with Adapter.

2) If it wasn't pushed, people wouldn't really try it out.

3) Mekle likes .NET too much.


You're really overstating things. We're not forcing anyone to move development studios. All of the languages you could develop in before TOD, you can develop in after TOD. Adapter requires compilation against version releases at the moment. Again, I don't think this is an onerous burden for anyone who's actively developing a plugin.

 

-----signature-----

I don't understand why everyone assumes that just because we make no promise that things won't change (and in fact say in advance that certain things will), that it means that everything will change and that nothing is stable.


for example:


Scroll Watcher 0.4.0.112, which was written in VB6, still works fine in the latest release (Decal 3 Alpha 5) and it hasn't needed any changes at all, not even recompiling, since it was updated for Alpha 3 in mid-October.


Insane-Bot Client 1.1.0.30 was also written in VB6. It only needed to be recompiled for Alpha 5; no source code changes were needed.


from release to release, the number of breaking changes is relatively small (and not all of them break for everyone). people who claim that they don't want to develop to a moving target either don't understand just how small the changes are from version to version, or they're just looking for an excuse not to update.


-ken

It would also help any developers to actually hang out near the SOURCE of the development in #ACDev and ask questions. You can ask some of the other devs like -Gol- who have been hanging out there... there is a lot that goes on in there which is important and really helps one understand the changes that are happening (or the lack of changes)