Hi! Great that there is some traffic again, I was afraid slide might be dying.
I added my 2 cents inline: > -----Original Message----- > From: Eric Johnson [mailto:[EMAIL PROTECTED] > Sent: Tuesday, September 13, 2005 23:46 > To: Slide Developers Mailing List > Subject: Re: project direction > > > Paul Hussein wrote: > > > Thanks for the comments > > > > As far as the client goes, I was thinking more of a client > that gives > > direct access like webdrive does, but with editing of acl's and > > searching maybe like google desktop, all built-in to the windows > > desktop, so you could map a network drive to a repoitory, > search it, > > click on a file and set the acl's. > > I don't know how many of us I speak for, Speaking for me! > but one reason for > working in > the Java world is to decouple us from the platform. But that is > slightly off topic. > > Back in Java land, a "client" might make sense, either as an > extension > to Commons-VFS (curious, that project already has webdav:// > support), or > say, as an extension to the Eclipse framework, and/or as a > NetBeans plugin. There is already a very good eclipse plugin: http://www.s-und-n.de/sunshine/ccos/de/produkte/english/webdavpilot Funnily, the newest version is only announced on the german page http://www.s-und-n.de/sunshine/ccos/de/produkte/webdavpilot, meaning version 0.3.3. This one "does" ACLs, you can also modify the directory structure and modify properties. The new version also does version diffs. Search is still only "planned" though. You can get it here: http://www.s-und-n.de/sunshine/download2/de/webdavpilot0.3.3.zip > > For one of our applications, I've done work using the Slide > client side > libraries to write a substitute for using a version control system. > Although I cannot contribute that code to a project (nor might anyone > want to look at it!), it does point the way to interesting > client side > applications. There are lots of ways to use a webdav server like slide, not only through the client api. However I think the reason why many people have discontinued working on slide and might have switched to jackrabbit is that slide is a big monster inside. I don't know to what extend this is so with jackrabbit, but that one is still growing, so people can have more influence on how to "do it right". In my opinion, people working on/with slide should probably focus on the usability aspect from the server side. Also, while slide is sort of a reference implementation, it does need to be fast, which is simply not the case right now. DASLs, the feature we use most, is slow as heck, even in the JDBCStore. I've added some features to spice up the lucene based dasls from the trunk a little and streamline the whole basicsearch stuff. I've already submitted a patch, but I have to clean it up a bit more, which is not so easy with my current workload... In any case, I think that since slide is pretty mature, we should focus on streamlining, meaning speed optimizations and configuration tools. It would be great to have a JMX console for the domain for hot configuration, some scripts or so to install slide, maybe bundle jetty in the source like cocoon does it for a quick test drive feature and maybe even as the main deployment platform, things like that. Clients are nice, but like people said before, they mostly exist. Another nice thing would be to be able to use slide to "webdav-enable" your own servlet application. This would be nice for people who are building a web application and would like to publish some info (like calendaring information via iCal) using webdav. This would mean a more flexible store implementation, and an API to add a store or a whole namespace on the fly. In general, I think slide needs a makeover. It seems very hard to maintain and to use. How about a major refactoring of the authentication / storage / locking / services components for version 3? I think they should be more separated and transparent, so that some layers can be skipped for applications that don't need them and so that programming in one layer becomes easier (e.g. we don't have to worry about locking anymore in the authentication layer). Right now, I see a lot of checkCredentials() and checkLocks() all over the place, along with many instances where SlideTokenWrappers are needed to either disable or enable locking or transaction management in some cases for some obscure reason(s). Well, enough said for today, back to work! :D Best regards, Max Pfingsthorn Hippo Oosteinde 11 1017WT Amsterdam The Netherlands Tel +31 (0)20 5224466 ------------------------------------------------------------- [EMAIL PROTECTED] / www.hippo.nl -------------------------------------------------------------- --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
