Sergiu Dumitriu wrote:

Vincent Massol wrote:

...

Hi,

As you know I like clean stuff (some could call it an obsession ;)). While browsing the Design/Idea spaces on dev.xwiki.org I noticed several pages containing stuff already implemented.

I'd like to propose the following strategy:

* Once a design has been implemented, document it (for example in the Modules space on code.xwiki.org) * When both the new design has been implemented and documented, delete the wiki page in the Design space.

Shouldn't we keep those for historical reasons, somewhere outside the Design space? How about DesignArchive? We could move a page there, and add at the end or at the top a conclusion, like "Implemented, see [the documentation page]", or "Deprecated, see [alternative design]".

+1 Thanks, Marius

...

For example I'd like to remove this page: http://dev.xwiki.org/xwiki/bin/view/Design/XWikiSyntaxMapping

I'd also like to document the xwiki-component module in Modules and then remove this page: http://dev.xwiki.org/xwiki/bin/view/Design/ArchitectureV2

This one should be kept IMO, since it contains more than just the description of the component architecture.

On Aug 17, 2009, at 2:35 PM, [Ricardo Rodriguez] Your EPEC Network ICT Team wrote:

Hi all,

Sergiu Dumitriu wrote:

...

Vincent Massol wrote:

...

Hi,

As you know I like clean stuff (some could call it an obsession ;)). While browsing the Design/Idea spaces on dev.xwiki.org I noticed several pages containing stuff already implemented.

I'd like to propose the following strategy:

* Once a design has been implemented, document it (for example in the Modules space on code.xwiki.org) * When both the new design has been implemented and documented, delete the wiki page in the Design space.

Shouldn't we keep those for historical reasons, somewhere outside the Design space? How about DesignArchive? We could move a page there, and add at the end or at the top a conclusion, like "Implemented, see [the documentation page]", or "Deprecated, see [alternative design]".

I've been supporting here a position where historical archive of any change in a given document is one of the main reasons to consider a wiki as the core of an information system. I would also like to see an alternative that doesn't imply to delete the seed document even though it is superseded by a new document.

This is about cleaning up. No system can go on indefinitely by keeping everything online. The strategy is simple: - you backup (as we do every day) and remove things that shouldn't be kept. They're still available on backup should they be needed one day. As for revisions we'll need a way in the future to be able to maintain that too since they grow quite large. I don't agree about absolutely keeping old things online when they're not needed anymore. If we had done this we would still live with some mess of xwiki.org as it was back in 2005-2006. The main point of a wiki is to *NOT* keep stale things by applying constant refactorings (which is what I suggest) so I definitely don't agree about using a wiki as an archival system (there are way better solutions for that) :) Thanks -Vincent

Cheers,

Ricardo

On Aug 18, 2009, at 1:01 AM, [Ricardo Rodriguez] Your EPEC Network ICT Team wrote:

Hi,

Vincent Massol wrote:

...

This is about cleaning up.

I am almost sure now: I've just jumped into the wrong thread :-( Sorry if I am adding nothing but noise.

...

No system can go on indefinitely by keeping everything online. The strategy is simple: - you backup (as we do every day) and remove things that shouldn't be kept. They're still available on backup should they be needed one day.

I do agree. But I've a huge problem there: how to decide which "things" should not be kept. Yes, they will be on backups, but how to be sure that "the system" will be able to find them when needed?

This is the usual question and the answer is usually that you need to keep a reader able to read them along with the data. Even more you need to maintain the reader over time to ensure it continues running on the hardware. Then how long the data is kept online completely depends on the domain and use case. In your case you could keep it 10 years online or more. Provided you have enough disk space and your DB is correctly indexed, that should work fine. For a documentation site it's a bit more problematic since the site needs to be relatively clean and when users go to it they should not feel it's a mess or a work in progress too much since otherwise this feeling will reflect on their feeling for the software. In the case at hand it's acceptable since it's on the dev wiki and not the main documentation wikis.

...

As for revisions we'll need a way in the future to be able to maintain that too since they grow quite large.

As devs say when you agree, +1 :-)

...

I don't agree about absolutely keeping old things online when they're not needed anymore. If we had done this we would still live with some mess of xwiki.org as it was back in 2005-2006.

Again, the same problem. How to decide that a "thing" is not needed? I'm afraid I must make an effort to present an example. A case study. I am thinking/working in a research environment. It is really hard to decide what could be deprecated at a given moment.

...

The main point of a wiki is to *NOT* keep stale things by applying constant refactorings (which is what I suggest) so I definitely don't agree about using a wiki as an archival system (there are way better solutions for that) :)

I agree! I don't think about a wiki (XWiki) as archival system. I think that I think (sic) about XWiki like a key piece in an information system helping to move textual information (prose explanation about data and thoughts) into structured documents, then into knowledge.

Perhaps documents can/must not last forever, but years should be considered as a regular duration for a document life cycle. During this time, it is possible that the size of modification won't be an issue: a single document won't growth forever, but the number of relations will do.

I've to read a much more about this!

Well I think we both agree: 1) wikis are tools that make it easy to keep stuff up to date and apply refactorings to content 2) in general it's bad form to loose content For 2) I was focusing on the content representing the solution in the design docs which I was proposing to drop from the design doc since I was proposing to move it elsewhere (in the main doc location). However I was indeed also suggesting to loose content (which was the comments in the design docs or some intermediary steps that allowed us to reach the final solution). However Sergiu didn't agree and I could understand why. So I backtracked and I've moved done design docs in a DesignArchive space. At some points it'll get removed but it'll probably be in a few years now. Thanks -Vincent

Thank you very much for your thoughts!

Ricardo

-- Ricardo Rodríguez Your EPEC Network ICT Team

hi On Wed, Aug 19, 2009 at 1:06 AM, [Ricardo Rodriguez] Your EPEC Network ICT Team <[email protected]> wrote:

Hi,

Vincent Massol wrote:

...

This is the usual question and the answer is usually that you need to keep a reader able to read them along with the data. Even more you need to maintain the reader over time to ensure it continues running on the hardware.

Then how long the data is kept online completely depends on the domain and use case. In your case you could keep it 10 years online or more. Provided you have enough disk space and your DB is correctly indexed, that should work fine.

I know this is a kind of "thread kidnapping", but here we have one our main concerns right now: Lucene search seems to have some issue that prevents it to work fine.

For instance. I get 70 hits (logged search) when looking for "Servlet"...

http://www.xwiki.org/xwiki/bin/view/Main/LuceneSearch?text=Servlet&x=0&y=0

Only a variable number of entries are shown in each page, but never the total number.

I think this is a major issue concerning how to retrieve stored contents. I don't know if it could be my Internet clients that are doing something wrong, but I keep getting this consistently.

What do you mean because when I always get the same number of hits with 30 hits per presentation page and 3 pages... From my point of view of XWiki.org user, sometimes, there are some strange results or no result but I think this is due to the fact that the site content has been changed or the server redeployed and lucene has not indexed everything yet or there has been a problem while indexing. I've already asked several times about it and each time, when people look at my problem, it's OK again because lucene periodically indexes as far as I remember :) But you may speak about something else...

...

For a documentation site it's a bit more problematic since the site needs to be relatively clean and when users go to it they should not feel it's a mess or a work in progress too much since otherwise this feeling will reflect on their feeling for the software. In the case at hand it's acceptable since it's on the dev wiki and not the main documentation wikis.

I can only agree with this. It seems to me that I am being a bit naive thinking about a unique repository of prose/data/information from which applying rules and policies will shown the right facet for each "type" of users.

I agree about the fact that clear documentations eases the access of new users to any new system/software.

...

Well I think we both agree: 1) wikis are tools that make it easy to keep stuff up to date and apply refactorings to content 2) in general it's bad form to loose content

For 2) I was focusing on the content representing the solution in the design docs which I was proposing to drop from the design doc since I was proposing to move it elsewhere (in the main doc location). However I was indeed also suggesting to loose content (which was the comments in the design docs or some intermediary steps that allowed us to reach the final solution). However Sergiu didn't agree and I could understand why. So I backtracked and I've moved done design docs in a DesignArchive space. At some points it'll get removed but it'll probably be in a few years now.

I do hope in a few years time we have the solution for to-be-deprecated contents!

What do you do today with your 10s of HardDrives with Mo, Go of deprecated data you don't even remember even if they could be interesting... The loss of digital information meaning also the loss of knowledge and maybe the loss of a part of our history will be one of the challenge in the future (this phrase is too serious so I put a smiley at the end :))... but there is also the case when some info keep going on internet and you would like it to be erased but it's almost impossible :):):)

Thanks for your time,

Ricardo

-- Ricardo Rodríguez Your EPEC Network ICT Team

_______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs

On Wed, Aug 19, 2009 at 9:18 AM, [Ricardo Rodriguez] Your EPEC Network ICT Team <[email protected]> wrote:

Hi!

...

...

Hi,

Vincent Massol wrote:

...

What do you mean because when I always get the same number of hits with 30 hits per presentation page and 3 pages... From my point of view of XWiki.org user, sometimes, there are some strange results or no result but I think this is due to the fact that the site content has been changed or the server redeployed and lucene has not indexed everything yet or there has been a problem while indexing. I've already asked several times about it and each time, when people look at my problem, it's OK again because lucene periodically indexes as far as I remember :) But you may speak about something else...

Sergiu has been recently dealing with this issue. Please, see this...

http://n2.nabble.com/Lucene-again-td3273364.html#a3273364

I'm getting now similar results with another searches. For instance, if I do a logged search looking for "Servlet", I get this...

http://xen.net/images/searchingServlet.jpg

Only four entries out of 30 in the first page of results. The second one shows only 3 and the third one, eight out of ten.

I am using Firefox 3.5.2 in a Mac OS X 10.5.8 box, and also get the same results using updated releases of Safari and OmniWeb. Cache is clean.

...

What do you do today with your 10s of HardDrives with Mo, Go of

deprecated

...

data you don't even remember even if they could be interesting... The loss of digital information meaning also the loss of knowledge and maybe the loss of a part of our history will be one of the challenge in the future (this phrase is too serious so I put a smiley at the end :))... but there is also the case when some info keep going on internet and you would like it to be erased but it's almost impossible :):):)

Gonna change the sentence in our home page right now!!! :-) Even though I do prefer to stick with this another one from you also with a smiley at the end...

Everything is possible... your imagination is the limit :)

Ricardo, I want to keep the copyright of my sentence and let Vincent with his sentence and I really disagree with him because my imagination has no limit :)

http://n2.nabble.com/Programming-Help-td509957.html#a17229158

After all, not all is lost in the net. Let's XWiki cope with the challenge!

I will be relatively far from computers for five days. But of course I will do as much as I can to follow all the interesting topics that are being discussed in the lists. Thanks for your time.

Cheers,

Ricardo

-- Ricardo Rodríguez Your EPEC Network ICT Team

_______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs

On Aug 17, 2009, at 11:21 AM, Sergiu Dumitriu wrote:

Vincent Massol wrote:

...

Hi,

As you know I like clean stuff (some could call it an obsession ;)). While browsing the Design/Idea spaces on dev.xwiki.org I noticed several pages containing stuff already implemented.

I'd like to propose the following strategy:

* Once a design has been implemented, document it (for example in the Modules space on code.xwiki.org) * When both the new design has been implemented and documented, delete the wiki page in the Design space.

Shouldn't we keep those for historical reasons, somewhere outside the Design space? How about DesignArchive? We could move a page there, and add at the end or at the top a conclusion, like "Implemented, see [the documentation page]", or "Deprecated, see [alternative design]".

I understand your point, which is to not loose the "discussion" that produced the designs, including comments from people, etc. BTW we already have a "Completed" state in Design. However you don't see it when you browse the design space using XEclipse or WebDAV so at the very minimum moving them out in a DesignArchive space would help. Ok I'll move them there for now even though the likelihood that we'll need them is probably low.

...

For example I'd like to remove this page: http://dev.xwiki.org/xwiki/bin/view/Design/XWikiSyntaxMapping

I'd also like to document the xwiki-component module in Modules and then remove this page: http://dev.xwiki.org/xwiki/bin/view/Design/ArchitectureV2

This one should be kept IMO, since it contains more than just the description of the component architecture.

Then this additional stuff needs to be documented properly somewhere IMO. Thanks for the feedback -Vincent