- From: Henry S. Thompson <ht@inf.ed.ac.uk>
- Date: Tue, 25 Jul 2006 17:27:07 +0100
- To: Sean Martin <sjmm@us.ibm.com>
- Cc: public-semweb-lifesci@w3.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Sean Martin writes: >> Well, either your scheme is intended to be dereferenceble, or it >> isn't. >> >> If it is, then instances are likely/virtually certain to contain some >> kind of named starting point, which needs to be looked up and >> resolved to an IP address start the dereferencing process. Domain >> names and DNS are by far the best available implementation of this >> step, with excellent performance, widespread deployment and >> considerable flexibility. > > As it is a URN, the starting point for dereferencing is urn.arpa. The > specification [1] details the use of the DDDS system (RFCs 3401-3405)which > uses the existing DNS system (for the very reasons you detail) but > maintains a level of abstraction between the authority name in the > identifier and the data service location that can provide a copy of what > was named, as is proper for URNs. So, register one of lsids.org, lsids.net, lsids.name or lsids.info, and use e.g. http://lsids.or/xxx instead of URN:LSID:xxx. Bingo -- no new tools required, works in all modern browsers :-). Implement as much or as little redirection, caching etc. as you wish in the server you run at lsids.info:80, just as you would using DDDS. >> >> as well as the only means by which one may >> >> retrieve it (the protocol, usually http, https or ftp). >> >> Not so. The URI RFC [1] makes clear that it is up to protocols to >> specify what URIs they interpret and how, not the other way around. >> It is entirely reasonable, and indeed expected, that new protocols may >> specify interpretations of 'old' URI schemes, including 'http'. >> >> >> The first question to ask yourself here is that when you are >> >> uniquely naming (in all of space and time!) a file/digital object >> >> which will be usefully copied far and wide, does it make sense to >> >> include as an integral part of that name the only protocol by which >> >> it can ever be accessed and the only place where one can find that >> >> copy? >> >> I hope the above clarify that this is not the case for names using the >> 'http' scheme. Indeed they are much more likely to do so for 'http' >> than for almost any other scheme. > > Assuming that a new http protocol replaces the existing one, how does this > change things? Sorry I wasn't clearer. Any new protocol of _any_ kind can specify how it handles 'http'-scheme URIs. I assumed you were worried about 'http'-scheme URIs going "out of date" somehow because 30 years from now http is dead and we're all using sdtp (super-duper transport protocol TM, Patent Pending :-). All I'm saying is that for sdtp to be a success, it will surely define what to do with an 'http'-scheme URI. > Surely the name is still tied to a single protocol (HTTP) > even if the underlying implementation of that protocol has changed? No, again, as the URI RFC makes clear, it's _protocols_ which define what they do with URIs of particular schemes, not the other way around. > LSIDs are independent of any particular transport protocol and > indeed already make use of any of the commonly used ones > simultaneously (ftp, http, SOAP, file:// etc). The thing to remember > here is that we are not thinking about URIs in the abstract here, > but rather a 'living, breathing system' intended for naming digital > objects that will be copied/archived far and wide. It was deemed > important to support as many mechanisms as possible (including > future ones) to support that copying/archiving process without > losing track of the unique name. So all LSID clients have to support all those protocols? Doesn't sound like a likely route to wide deployment. . . Or are you proxying all requests through a few central servers, who choose what protocols to use for the initial fetch? If so, no problem doing that with 'http'-scheme URIs either. . . >> >> Unfortunately when it comes to URL?s there is no way to know >> >> that what is served one day will be served out the next simply >> >> by looking at the URL string. There is no social convention or >> >> technical contract to support the behavior that would be >> >> required. > >> >> True for some 'http' URIs, false for others. The owners of a group of >> names, whether they use 'http' or not, are responsible for >> documenting, implementing and enforcing usage conventions. I >> absolutely agree that for your purposes you need to take this very >> seriously, but using 'http' doesn't make this any harder (or, of >> course, any easier). >> > > I am not sure that I can agree with you on this point. How does one go > about differentiating between one http:// URI and another programmatically > for the purposes of knowing what its conventions are? As opposed to using > something else which only has one established convention? See above suggestion wrt http://lsids.org/ -- you own that domain, you set the conventions/policies/etc. > [1] http://www.omg.org/cgi-bin/doc?dtc/04-05-01 > > -- > Sean Martin > IBM Corp. > - -- Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh Half-time member of W3C Team 2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440 Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk URL: http://www.ltg.ed.ac.uk/~ht/ [mail really from me _always_ has this .sig -- mail without it is forged spam] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2.1 (GNU/Linux) iD8DBQFExkZbkjnJixAXWBoRAlWAAKCEBW+ClErSYh6+p1Haxhoi+3n0UgCeK6xP tiF46Ziales3L9xL2omz+r4= =LfH/ -----END PGP SIGNATURE-----
Received on Tuesday, 25 July 2006 16:27:30 UTC