From: Max V. <vo...@fz...> - 2007-07-31 19:25:27
|
L> If you use IBAW, also do it on the wiki page L> http://nepomuk.semanticdesktop.org/xwiki/bin/view/Main/LocalId Ok, I reformatted the page a little. L> or inside the RFC draft: L> http://svn.nepomuk.semanticdesktop.org/repos/trunk/paper/2007/semanticde= sktop_urischeme/RFC_semdesk_urischeme.txt would be much nicer to do this in the wiki as well. L> I changed the document repeatedly based on your feedback, discussing=20 L> only via e-mail will get us nowhere, L> as the version of the text in this mail does not reflect the version on= =20 L> the wiki or the draft. ah, ok. L> answers inline, and on the wiki and in the draft. ok, I re-read the draft and commented on the wiki page. >> L> I also think that creating this standard within Nepomuk, and later >> L> asking W3C or OASIS to bless it will not work. >> I think this might very well work. L> no, because: L> * A- nobody of us has deep understanding of desktop search engines,=20 L> desktop operating systems, and especially the plans of desktop operating= =20 L> system vendors. at W3C they are members and we have the chance to get=20 L> the right people into the standard. Namely, people that have the same=20 L> problem and have an industrial motivation to standardize it. L> * B - once nepomuk has written a standard, it cannot be standardized as= =20 L> W3C or ISO/OASIS standard as this is only a "suggestion" that will be=20 L> discussed in a second WG again, so why on earth have two discussions if= =20 L> you can start with the right one right away? L> * C - if the desktop OS people can be motivated to join a W3C WG on this= =20 L> URI standard, they might even implement it. ok, I don't really care - I am totally in favour to involve a broader gro= up in the beginning. You convinced me here. >> We have to consider the pro and con between: >> - adding a new URI scheme >> - adding something to DNS, i.e. it is not by accident that most OS al= low very >> easily to tweak DNS resolving. Using http://somename.localhost/ where = somename >> identifies the user might not bew such a bad solution. Let's kill= it with >> arguments, not emotions :-) >> =20 L> we don't need a networked solution for non-networked resources. L> This is about DESKTOP resources resolved on one DESKTOP. L> for networked retrieval, there is HTTP and the nepomuk p2p-semantic=20 L> component. L> Also, networked retrieval can be done with PIMO as a mid-layer. ok, the argument is "we don't need inter-desktop-resolving", agree. L> This is also not about identifying concept from PIMO or minting URIs for= =20 L> user-annotations, L> this is handled in Nepomuk and may be HTTP. L> At this point, its really only about infomration resources handled by=20 L> the operating system and existing applicaitons, extracted by DataWrapper= s. This was not clear enough for me, thanks for remarking. I added this to the wiki page. L> If you take HTTP as the scheme for desktop addresses, and you call=20 L> "OPEN" on the URI, like clicking it in the browser, the standard HTTP=20 L> protocol handler of the operating system will be called. This is good,= =20 L> because its a standard.* L> *We would destroy the standard by tweaking it beyond recognition for the= =20 L> desktop cause. It depends what tweaking we need. There is no OPEN in http. L> again: an OPEN command to a HTTP-URI will open a web-browser, not my PIM= =20 L> application. L> thats why you have so many uri schemes (magnet, kde, akonadi) Ok, requirment: "launch application with given object" - indeed not clear how to realize this with HTTP. Maybe: URL?action=3Dopen Its not impossible. L> I would appreciate we invest time to learn how the existing desktop=20 L> identification schemes work and understand them, and integrate them,=20 L> instead of continuing on the http track here. But you have no clear argument for this. Maybe simlpy because you found n= o way to implement your requirments with HTTP - so there should be at least one= req. which has only a very ugly HTTP impl. L> Now comes a lot of pro-http, which again I don't want to discuss,=20 L> because it does not help much. Thats a pretty strict statement without a clear argument behind it. >> L> And >> L> HTTP uris are not retrievable for desktop pcs. >> They are, if they are on localhost. >> >> L> Even with the hacks we >> L> thought of, it is not clean. The HTTP port for a desktop may be block= ed >> true! >> >> L> or multiple users on one machine may collide ports. >> aha, a deployment problem: this would mean, one cannot install several = NEPOMUKs >> on one machine. Really? How do other http-server application handle th= is, e.g. >> Google Desktop search? Eclipse help system? >> =20 >> L> Also, opening an http server on the desktop may be a security risk in= # >> L> many environments. >> On the other hand, its a risk that can be controlled by accepting on= ly local >> connections or whatever. Traffic can be monitored. >> >> L> And, there is the file:// uri scheme >> L> standardized by IETF that shows how to do it. >> I don'T see the relation to HTTP. >> >> L> http is not the silver bullet on the desktop, I have to admit. >> Agree. >> >> L> 1. The "semdesk" URI scheme >> L> Semantic Desktop URIs have several parts, including an application sp= ecific part, >> L> hostname information and a path within the application's namespace. >> >From this i derive the requirement to adress: >> >> machine x user x application x application-object >> >> L> The name "semdesk" stems from the requirement to identify resources f= or >> L> further annotations using the RDF technology. On desktops that are not >> L> aware of annotations or semantic databases, a way of uniquely identif= ying=20 >> L> resources is also handy, thus it is not required to run RDF applicati= ons=20 >> L> to use the semdesk URI scheme.=20 >> Good! >> >> L> The scheme is intended to be used by desktop search engines to identi= fy=20 >> L> resources from desktop applications. This allows resources to be=20 >> L> identified independent of implementation. >> Not *only by dekstop search. >> =20 L> point there, please edit the file to fix it: L> http://svn.nepomuk.semanticdesktop.org/repos/trunk/paper/2007/semanticde= sktop_urischeme/RFC_semdesk_urischeme.txt Let's discuss on the wiki page first between HTTP and new URI scheme= . You intention is that the OS should handle the new scheme, right? >> >From an implementation standpoint, how can I invoke this protocol via >> - Java >> - Ruby >> - C >> - C# >> - Python >> - shell scripts >> - ... >> =20 L> its platform dependent, as said. L> They call the normal operating system's "START URI" command. This implies that all OS which run semdesktop stuff have to change - or ha= ve to be changed by us. This is just something to consider. >> To clarify, I really see the problem you try to solve, Leo. I don't kno= w what a >> good solution is. I guess, we should try to list the requirements, fil= ter them >> and think of the best solution. I wouldn't kick HTTP to easily, mainly= because >> implementation can be very easy (at least it *seems* right now). >> =20 L> no, its not easy, because we have to re-impleemnt http for the desktop= =20 L> avoiding all the pitfalls mentioned above. L> We should standardize semdesk: first, and then you can give it a try to= =20 L> port it to a HTTP hack, but not the other way round. Again, no argument why. I don't believe HTTP is better, but neither see I clearly semdesk as the b= etter solution. It depends on what we want to achieve and what amount= of implementation effort has to be provided by whom. Maybe semdesk is the way = to go because its nucer or cheaper - I just don't see that yet. Kind Regards, Max -- Max V=F6lkel office: +49 721 9654-854 http://Xam.de mobile: +49 171 8359678 FZI Forschungszentrum Informatik http://www.FZI.de an der Universit=E4t Karlsruhe telephone: +49-721-9654-0 Haid-und-Neu-Str. 10-14 fax: +49-721-9654-959 D-76131 Karlsruhe Stiftung des b=FCrgerlichen Rechts. Az: 14-0563.1 Regierungspr=E4sidium Kar= lsruhe. Vorstand: Prof. Dr.-Ing. R=FCdiger Dillmann, Dipl. Wi.-Ing. Michael Flor, Prof. Dr. Dr.-Ing. Jivka Ovtcharova, Prof. Dr. rer. nat. Rudi Studer. Vorsitzender des Kuratoriums: Ministerialdirigent G=FCnther Le=DFnerkraus. |