Hello Daniel, Thanks a lot for your reply. Please find my comments inline.
> -----Original Message----- > From: Daniel Florey [mailto:[EMAIL PROTECTED] > Sent: Tuesday, March 30, 2004 3:08 PM > To: Slide Users Mailing List > Subject: Re: Which API should I use for a web app? > > Yes, Slide is an abstract content repository but it depends > on the kind > of application you want to build on top of it, if it is really usable. > What I was talking about was a portal like webapplication > using some of > the content displayed to the user by using slide. If you are > thinking of > a totally different app, the performance problems might not occur. > Let's say you are thinking of a webapp that has several > JSP-pages that > work without retrieving data from slide while generating > output and you > have a download area where users can download documents it > might be ok. > But if you think of web pages that contain content that is stored in > slide and will be retrieved while generating the page it will > be really > slow. The API you use will not make a very big difference as the > performance problems occur inside the slide kernel > (permission checking > etc.) > If the performance impact occurs inside the Slide Kernel, then how would it be different when accessed via a Web-Client as opposed to a Desktop-Client? The only difference is: in case of a Web-Client I would use some JSP/XSP and in case of a Desktop-Client I would use some Web-Service. With both the JSP/XSP and Web-Services internally using the Slide Client LIB to access Slide's WebDAV Service. In your example above, when you say "retrieving data from Slide", do you imply retrieving actual Content or retrieving any information: MetaData or Content? E.G. 1. if the client(Web/Desktop) wants to view all the sub-folders inside a folder, a Slide WebDAV <ls> command would be issued and the results would be wrapped in respective format and returned to the client. E.G. 2. if the client(Web/Desktop) wants to download all the files in a folder, a recursive Slide WebDAV <Get> command would be issued and the results would be wrapped in respective format and returned to the client. Will there be performance impact in both cases or only case 2 (large information download)? Sorry for going into so much details... But I really did not understand the difference between a Web-Client and a Desktop-Client. The security and lock checks would be required in both cases. And most likely both cases would be communicating with server over http. > >When you mention "background-building of webpages"... are > you referring to > >the fact that Slide communicates over the WebDAV(HTTP > extension) protocol > >and by that fact it would be required to return a webpage in > response to any > >request? > >If yes, would that mean that the performance impact is due to the > >communication layer between a WebApp and Slide? In other words, the > >performance impact can be attributed to using Slide's Client > APIs inside our > >WebApp. And that could be avoided if we access Slide APIs > directly from > >inside our WebApp. Is this interpretation correct? > > > > > No, the api makes no big difference. You should use the webdav lib or > wvcm to access slide, otherwise your app is bound to the same vm. When using the Client LIB, wouldn't there be a big performance difference due to the additional HTTP communication layer introduced(I am referring to using Client LIB from inside a JSP or Web-Service on the Server side)? (Having the Slide APIs run in the same VM is acceptable). However I think not having a clean separation between the WebDAV and the Slide API layer would mandate the use of the Client or WVCM libraries. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
