I feel all of those user frustrations. I'd be happy to accept any help you can offer on these projects.
Every day I work on NAnt/Contrib I felt those issues and work towards getting them solved. No body will use your product if you can't install it and have it running in minutes. We put in place nightly builds to help out with some of these issues. It really shouldn't be that complicated to get working binaries; I'm firmly of the belief that you shouldn't need to recompile to use software (esp. when it is a build tool.). The problems is that we weren't able to get things done fast enough. There aren't enough hours in the day, and people on the project, to do what we could see was needed. As soon as it seems like we were close, we made larger steps to make things more manageable and better for the future. Sometimes if you want stability you need to take a few steps back on the features and live with those consequences. We are working on pre-release software (and in the case of NAntContrib there has never even been a beta release) in an free and open-source project. It comes down to the fact that nobody is paid, or made, to work on the project. I'm not slighted by your comments or email. I'm impressed that you took the time to write it. But let me say that you have hit on my biggest concerns with everything we are doing. Getting it to a state that I can pass it out as shrink-wrapped and ready to use. One of the goals for NAntContrib is to get it out with a version of NAnt so that everything works as a single product. It shouldn't matter what tasks you are using, or where you get them from, as long as they work for you. ----- Original Message ----- From: "Jim Sokoloff" <[EMAIL PROTECTED]> To: "Ian MacLean" <[EMAIL PROTECTED]> Cc: "Farris, Keith" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]> Sent: Friday, June 27, 2003 2:37 PM Subject: Re: [NAntC-Dev] User Frustration > > what are your particular grievences ? NantContrib needs some love - no > > doubt but nant itself is fairly stable right now - or is that not your > > experience ? > > I'll chime in with my experiences, even though I'm not Keith. :) > > I think the frustrations being voiced are that (at least during > the re-org) the software doesn't come build right out of > the box, and once the compile error/warnings are corrected, the > nantcontrib tasks (vssget at least) don't work, giving an error > message "task not found" or somesuch. As someone unfamiliar > with the code, it's a somewhat daunting task to fix. > > For my own usage, I backed off to 0.82 and a matching version > of NantC, but that meant removing a bunch of code that had > relied of 0.83-dev features in nant. I had to give those up > because I couldn't get a working 0.83 with a <vssget> task. > > > By "the software", I mean "nant and nantcontrib" not "nant". > To my mind, end users will never care whether <gac> <vssget> > <csc> <timestamp> or anything else come from nant or > nantcontrib; to them (particularly to those who don't build > the software but just use it), it's all just one piece of > software. > > I'm in a position where I need to recommend a build system > to my team (of >20 developers). I'm the new guy at this > company, so I don't want to blow my first major decision. :) > > I asked myself: > Is Nant/NantC stable enough for me to recommend to my team? > Is it likely to continue to evolve over time, while remaining > reasonably stable? > > At the *INSTANT* in time that I started looking at it (which > happened to be right around the time of the source re-org), > the answers appeared to be "NO!". Bad timing to be sure, but > the tree stayed broken for weeks (and I suspect is still > broken). > > When users are inquiring "Why couldn't there just be one > tree?" I think that comes from the idea "If there were only > one tree, it wouldn't stay broken for so long." (Near as I > can tell, it's been about 3 weeks.) > > As an end user, I don't care if there are one or two trees > (and I see a lot of benefit to a more "lightweight" NantC > tree to let more people contribute). After all, when > talking about the stable (working) trees, the build is > quite easy and the end product looks like one product > anyway. The problem is when something gets broken, it will > stay broken longer, because a nant developer who changes > something won't necessarily make the corresponding updates > in contrib. > > Thanks for reading so far. > > To remove myself from the category of "idle complainer", > I'd like to volunteer to help with nant[C], in particular > in writing some automated self-tests for it so we can > notice the breakages earlier and present a more stable > product to all users. > > What's stopping me? Nothing; I just need to clear a few > things off my plate at work and dive in. > > ---Jim > > PS: Because no one knows me from Adam, I'm a little > hesitant to even write this email. I know that no one > owes me anything; my offer to help fix what I see as > one of the current deficiencies is not intended as a > slight. > > PPS: In a past life, I led a team that built our own > in-house proprietary automated build/version/test/package > system. Boy what I wouldn't give to have those 4-5 man- > years back to apply to the nant project. :) ------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01 _______________________________________________ NAntContrib-Developer mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/nantcontrib-developer
