On 6 October 2010 14:16, Henning Thielemann wrote:

 I'm generally not glad that some people rearrange existing structure and expect that all of the affected authors follow. It's already tedious to catch up with the yearly changes in GHC's package and other base packages (e.g. transformers recently), and annoying when people propose to mark packages as "inactive" or "unmaintained" in Hackage whenever the package authors did not update their packages so far (and certainly lose compatibility to older 'base' versions this way).  I would be glad if there is no further action to be taken for package authors, who added their packages somewhen in the past and don't see a reason to regularly check whether their packages are still listed in the Wiki, without being marked "inactive" or so. If you think the re-structuring is necessary, then at least ask the maintainers, whether they still maintain their packages, or just sort the packages according to the degree of activity you assume, but stay away from categorizing the packages in "active" and "inactive" based on speculation.

Okay, don't worry about it, I'll do it!

On 6 October 2010 16:33, Henning Thielemann wrote:

How about mailing to the package maintainers in order to inform they, that the Web application list on the Wiki has changed? I'm afraid not all authors follow haskell-cafe or haskell-web.

I could send out a bulk mail requesting authors to have a look and help out improve this part of the wiki.

httpd-shed seems to be missing in Servers. I also like to see HWS mentioned in Servers as it is the ancestor of some Haskell Web Server projects (WASH-wsp, MoHWS, and what was the name of the CGI thing?). For me the Wiki is not only a place that describes cutting edge software but also a place to help understand how things evolved. HWS is still interesting, because its quite basic, so it's still a good start if you like to program your own server. It is not necessary to be maintained in order to be interesting.

I hadn't heard of httpd-shed. Will you add it to Servers? I think a page about HWS would also be good that shows the history of it and derived projects, if you feel like writing it! I also agree that even the simple examples like HWS are interesting, like CGI; I cleaned up the old CGI article: http://www.haskell.org/haskellwiki/Practical_web_programming_in_Haskell I condensed it visually, and updated links to be more within the wiki and separated, e.g. this page http://www.haskell.org/haskellwiki/Web/Literature/Static_linking because it's useful as a general idea and not focused on CGI specifically. It's definitely not my intention to discard useful information, just to make it more accessible and remove misleading data. Regarding the active/inactive, I think it's a good idea to separate what we know to be actively maintaned -- i.e., * what people are using, * what still has someone maintaining it, * what actually still *compiles*. Here's my reasoning, there are three uses of listing frameworks on the wiki: 1) People looking to survey what's currently available and stable -- i.e. what's alive? 2) People looking to try out Haskell web programming, who want something that they know will have some support and be current, therefore easy. 3) People who are serious about web development and want to survey the whole existing landscape. (1) and (2) don't care or want to have to sift through or waste time on libraries that don't work, or might not work. (3) has the motivation to sift through everything and they want to see the history of everything too. These people too will want to know what's current and working, I think. In this sense I think we are optimising access to the information. What do you think?

On Wed, Oct 6, 2010 at 2:16 PM, Henning Thielemann wrote:

On Wed, 6 Oct 2010, Christopher Done wrote:

...

On 6 October 2010 12:47, Henning Thielemann wrote:

...

I for instance use http-shed and mohws all the time. They do what they shall do for me. I maintain mohws

Please move the ones you use and maintain to the active list!

 I'm generally not glad that some people rearrange existing structure and expect that all of the affected authors follow. It's already tedious to catch up with the yearly changes in GHC's package and other base packages (e.g. transformers recently), and annoying when people propose to mark packages as "inactive" or "unmaintained" in Hackage whenever the package authors did not update their packages so far (and certainly lose compatibility to older 'base' versions this way).  I would be glad if there is no further action to be taken for package authors, who added their packages somewhen in the past and don't see a reason to regularly check whether their packages are still listed in the Wiki, without being marked "inactive" or so. If you think the re-structuring is necessary, then at least ask the maintainers, whether they still maintain their packages, or just sort the packages according to the degree of activity you assume, but stay away from categorizing the packages in "active" and "inactive" based on speculation.

I agree that it can be tedious to keep up with these changes, but the alternative is stagnation. Just a few years ago, I was in the place of the newbie staring at the wiki pages talking about all the wonderful ways of combining the CGI monad with fastcgi and xhtml and combinators, and something about monad transformer stacks (which I'd never even heard of). If I remember correctly, I gave up on Haskell for a month or so after the intimidation that kind of page introduces. Chris has done an amazing job here of cleaning up content and making it approachable by new users. I think that should be the main purpose of the wiki. If you want to have some documentation that no one else can edit, you can put it on your own site. That's what I've done for Yesod, and appears to be the approach of most of the other actively developed projects out there. It's true that in such a large reworking as Chris has undertaken there will be some accidental miscategorizations, but the alternate you mention (contacting each author before moving an article) is simply untenable: it's difficult to track people down sometimes, it takes a long time to get a response, etc. I'd much rather have a very clean looking wiki page that's missing a few packages than the jumble of confusion we had before hand. Michael