Re: [gmx-users] Mailing List Search on the new website
Hi Mark and David, I agree with the points you make. The change was rather abrupt but in the long term I believe the dekiwiki will be a much better and useful platform. The dekiwiki has been initially developed on top of mediawiki with the aim to make it more user friendly to modify content, e.g. using a WYSIWYG editor etc. It has very powerful scripting capabilities (using dekiscript) and a lot of useful extensions that make it possible to mash-up all sorts of content. You can easily embed videos and RSS feeds, view PDF/DOC/XLS/PS/RTF documets, SVG images or interface with tweeter/linkedin/skype(?) etc. It is basically mediawiki on steroids! Not that all of this is needed for this website but providing a rich platform will give the Gromacs community much more flexibility in sharing knowledge and experience. Tutorial videos, RSS feeds to sites of interest, current weather in Stockholm, etc. can enrich the site by a lot. I would have approached this process differently. The MediaWiki content was intrinsically free-form, and now we're apparently trying to shoehorn it into something hierarchical. This process converts hundreds of links that are now defunct because they were created as relative links to targets in a free-form database, and have now been automatically converted to what look like hard links to non-existent targets. The correct targets are presumably placed somewhere in a fairly arbitrary hierarchy (e.g. what's the difference between "Best Practices" and "How-tos"?), which anyone wanting to spend the time to fix now has to memorise. These categories existed in the previous wiki, but nobody adding content or links was forced to use or know them. That's a big change in philosophy. In fact, dekiwiki can be very easily set up as a graph/free-form style wiki. One just has to disable the navigation pane, one click away. Then you have only one page from which you can access the rest of the site. Also, in a wiki way, if you try to access a non-existing page you are automatically presented with the editor to add new page to the top level. Hyperlinks to other pages on the site are also relative as in a standard wiki, so even if you rename a page or move it to a different level in the tree (if you are using one) all links pointing to it will be able to find it. I've spent a lot of time over years creating and maintaining a large slab of this content (with other notable contributions from quite a few people - Dallas Warren in particular) and feel that this new imported format greatly reduces the utility of the information. While some curated hierarchical indexes are good things for helping to find content, I think preserving the underlying lack of hierarchy is a much better approach to preserving the usefulness of the content. The alternative is appointing a curator and having them spend lots of time organizing and fixing links. Given the slow rate of transition of GROMACS web content from old to new forms, I think we should be very keen not to make more centralized work, and to preserve the goodwill of previous "external" volunteers :-) I can easily remove the left pane and leave only a flat graph structure if that's better. It wasn't easy to categories all the info from the previous site but there is some inherent hierarchy in the type of content we have on the site. It is possible to combine both styles. How about keeping the tree and put easily "categorizable" content into the appropriate level? Then, have a top level "Others" (or something like that) where you can dump any page that doesn't fit the structure, and include link to it from another place. Same would hold for levels like "How-tos" and "Programming Guide". They can have 100s of sub-pages but you don't need to scroll down to find something. Put a new page there along with a link to it at some other page. Last but not least, there is a search form:) I'd much rather see the Documentation section (http://www.gromacs.org/Documentation) have links to * the paper manuals, * either a compatible Wiki, or a flat wiki-style import of the previous wiki (so that automated link conversion can just work - even if it's done with a simple perl script!), * some new and updated index pages for that previous wiki, * and if necessary, any other non-wiki content. Dear Mark, you raise many good points. A lot of the strong points of the simple wiki have been lost in the transition, I guess we have been misled by the wiki in the name Dekiwiki. Dekiwiki has some important advantages over mediawiki, like security (apparently the oldwiki was abused by spammers recently, this is why it was shutdown), and the ease for adding images and documents (mdp files :)). Whether dekiwki was the right choice for the job I am not sure. Overall it is easy to conclude that the move was done too hastily, seeing that the new website is still not anywhere close as informative as the two previous
Re: [gmx-users] Mailing List Search on the new website
Mark Abraham wrote: Rossen Apostolov wrote: Hi, It is now possible to search the mailing list archives from the new website: http://www.gromacs.org/Support/Mailing_Lists/Search . There is a link to it at the "Quick Links" section on the front page also. Currently: * Only the [gmx-users] mailng list is being indexed. * The index is updated every 2 hours. * The search input is not parsed, rather search is done only for the exact string. * The results are displayed in descending order of the date. You can change the sorting by clicking on the column names. * For some postings, the contents section in the table is empty but following the link will display the full contents. * It is likely that the interface will be implemented into the website differently. That's great. Hopefully we'll get more content and functionality soon too! Also, almost all of the information from the old wiki/www sites is now transfered and incorporated, although many pages contain broken links to them. They will be fixed one by one, but everyone is welcome to update them when you come across such. I would have approached this process differently. The MediaWiki content was intrinsically free-form, and now we're apparently trying to shoehorn it into something hierarchical. This process converts hundreds of links that are now defunct because they were created as relative links to targets in a free-form database, and have now been automatically converted to what look like hard links to non-existent targets. The correct targets are presumably placed somewhere in a fairly arbitrary hierarchy (e.g. what's the difference between "Best Practices" and "How-tos"?), which anyone wanting to spend the time to fix now has to memorise. These categories existed in the previous wiki, but nobody adding content or links was forced to use or know them. That's a big change in philosophy. I've spent a lot of time over years creating and maintaining a large slab of this content (with other notable contributions from quite a few people - Dallas Warren in particular) and feel that this new imported format greatly reduces the utility of the information. While some curated hierarchical indexes are good things for helping to find content, I think preserving the underlying lack of hierarchy is a much better approach to preserving the usefulness of the content. The alternative is appointing a curator and having them spend lots of time organizing and fixing links. Given the slow rate of transition of GROMACS web content from old to new forms, I think we should be very keen not to make more centralized work, and to preserve the goodwill of previous "external" volunteers :-) I'd much rather see the Documentation section (http://www.gromacs.org/Documentation) have links to * the paper manuals, * either a compatible Wiki, or a flat wiki-style import of the previous wiki (so that automated link conversion can just work - even if it's done with a simple perl script!), * some new and updated index pages for that previous wiki, * and if necessary, any other non-wiki content. Dear Mark, you raise many good points. A lot of the strong points of the simple wiki have been lost in the transition, I guess we have been misled by the wiki in the name Dekiwiki. Dekiwiki has some important advantages over mediawiki, like security (apparently the oldwiki was abused by spammers recently, this is why it was shutdown), and the ease for adding images and documents (mdp files :)). Whether dekiwki was the right choice for the job I am not sure. Overall it is easy to conclude that the move was done too hastily, seeing that the new website is still not anywhere close as informative as the two previous websites were. Rossen has done a lot of hard work on the website (e.g. the mailing list search), but there are some fundamental flaws like the lack of indentation in the menu structure, and the excessive amount of menus which apparently are difficult to change. You correctly identify the most important point that is missing: the self-organization that is inherent in the mediawiki format. I wonder whether it is impossible to get that functionality on dekiwiki (without that, one can not call it a wiki really...). The goodwill of external volunteers is another point that should not be overlooked. GROMACS is by now a community effort with many developers and contributor, big and small, scattered around the globe. To make the new website a success we do need many contributors, and we should try not to alienate contributors by such abrupt changes. Cheers, Mark ___ gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post
Re: [gmx-users] Mailing List Search on the new website
Rossen Apostolov wrote: Hi, It is now possible to search the mailing list archives from the new website: http://www.gromacs.org/Support/Mailing_Lists/Search . There is a link to it at the "Quick Links" section on the front page also. Currently: * Only the [gmx-users] mailng list is being indexed. * The index is updated every 2 hours. * The search input is not parsed, rather search is done only for the exact string. * The results are displayed in descending order of the date. You can change the sorting by clicking on the column names. * For some postings, the contents section in the table is empty but following the link will display the full contents. * It is likely that the interface will be implemented into the website differently. That's great. Hopefully we'll get more content and functionality soon too! Also, almost all of the information from the old wiki/www sites is now transfered and incorporated, although many pages contain broken links to them. They will be fixed one by one, but everyone is welcome to update them when you come across such. I would have approached this process differently. The MediaWiki content was intrinsically free-form, and now we're apparently trying to shoehorn it into something hierarchical. This process converts hundreds of links that are now defunct because they were created as relative links to targets in a free-form database, and have now been automatically converted to what look like hard links to non-existent targets. The correct targets are presumably placed somewhere in a fairly arbitrary hierarchy (e.g. what's the difference between "Best Practices" and "How-tos"?), which anyone wanting to spend the time to fix now has to memorise. These categories existed in the previous wiki, but nobody adding content or links was forced to use or know them. That's a big change in philosophy. I've spent a lot of time over years creating and maintaining a large slab of this content (with other notable contributions from quite a few people - Dallas Warren in particular) and feel that this new imported format greatly reduces the utility of the information. While some curated hierarchical indexes are good things for helping to find content, I think preserving the underlying lack of hierarchy is a much better approach to preserving the usefulness of the content. The alternative is appointing a curator and having them spend lots of time organizing and fixing links. Given the slow rate of transition of GROMACS web content from old to new forms, I think we should be very keen not to make more centralized work, and to preserve the goodwill of previous "external" volunteers :-) I'd much rather see the Documentation section (http://www.gromacs.org/Documentation) have links to * the paper manuals, * either a compatible Wiki, or a flat wiki-style import of the previous wiki (so that automated link conversion can just work - even if it's done with a simple perl script!), * some new and updated index pages for that previous wiki, * and if necessary, any other non-wiki content. Cheers, Mark ___ gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/mailing_lists/users.php
[gmx-users] Mailing List Search on the new website
Hi, It is now possible to search the mailing list archives from the new website: http://www.gromacs.org/Support/Mailing_Lists/Search . There is a link to it at the "Quick Links" section on the front page also. Currently: * Only the [gmx-users] mailng list is being indexed. * The index is updated every 2 hours. * The search input is not parsed, rather search is done only for the exact string. * The results are displayed in descending order of the date. You can change the sorting by clicking on the column names. * For some postings, the contents section in the table is empty but following the link will display the full contents. * It is likely that the interface will be implemented into the website differently. Also, almost all of the information from the old wiki/www sites is now transfered and incorporated, although many pages contain broken links to them. They will be fixed one by one, but everyone is welcome to update them when you come across such. Rossen ___ gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/mailing_lists/users.php
