Re: URL +1, LSID -1

IMHO "best practice" would be to use LSID client software that did this:

>   - otherwise, use DDDS or SRV: contact the authority via DNS
> protocol, get location of resolver

The HTTP proxies are really there so people can play with these  
things in browsers if they don't have support in their browser (e.g.,  
FireFox plugin or LSID LaunchPad for IE 6). They have no claims to  
stability or permanence.

You can NOT assume this:

>   - otherwise, maybe the authority itself has a resolver at
> http://foo.bar:8080/authority where foo.bar = the authority name
> (thanks Mark W)

Not every authority is on port 8080, nor does the authority resolver  
have to have the same name as the authority in the LSID. This is why   
you really want to look up the SRV record then pull off the WSDL file  
describing the authority.

So, you need software that resolves the LSID and retrieves the  
associated data or metadata. There are several kits that will do this  
right now (the SourceForge site has Java and Perl code, and I have a  
PHP client kicking around). LSIDs can be a bit complicated, but at  
least there is a very explicit protocol for getting the metadata (you  
don't have to guess, it tells you exactly how to retrieve it).

>   - otherwise, if the authority is offline, repurposed, or hostile,
> start using google, email, your rolodex, and your librarian to track
> down a resolver that knows about the LSID

I don't understand this bit:

> track
> down a resolver that knows about the LSID

This is about persistence, so isn't this really about caching, not  
resolving?  Of course, the aspiration is that the LSID authority will  
always be online ;-)   Ideally, people serving LSIDs would care as  
much about making them persistent as organisations like CrossRef care  
about ensuring that DOIs are always resolvable.

Rod


On 12 Jul 2007, at 17:04, Jonathan Rees wrote:

> I must have misspoken. What I meant is this.  Suppose I have an LSID
> (used as a URI sensu semantic web) that I got from email or a
> scientific publication, and I want to know more about the resource
> that it denotes. That is, I want some metadata for the resource, at
> the very least. My understanding is that to do this, you're supposed
> to talk to an LSID resolution service. In order to do that, you need
> to be acquainted with such a service.
>
> I'm not that interested in what is allowed or required; I'm interested
> in "best practices" for semantic web applications, and in particular
> the automated case, not browsers.
>
> Correct me if I'm wrong: Best practice is:
>   - if you have a list of LSID resolvers, or maybe HTTP bridges to
> some LSID resolvers, use those; maybe you got them from a friend, or
> they were wired into some software distribution(s)
>   - otherwise, use DDDS or SRV: contact the authority via DNS
> protocol, get location of resolver
>   - otherwise, maybe the authority itself has a resolver at
> http://foo.bar:8080/authority where foo.bar = the authority name
> (thanks Mark W)
>   - otherwise, if the authority is offline, repurposed, or hostile,
> start using google, email, your rolodex, and your librarian to track
> down a resolver that knows about the LSID
>
> This is similar to what I'd do for an HTTP URI: prior list of SPARQL
> endpoints or caches of metadata-carrying files, then DNS/HTTP to
> contact the named server (allowing for redirection, then using the
> current awful bag of hacks to guess where the metadata is), then
> wayback machine or Google cache (no LSID equivalent yet?), then pull
> out the big guns.
>
> It would be nice if the process for getting metadata (and data!) for
> resources identified by HTTP URIs were more systematic. The proposals
> I've heard are all awful, but I think it's a good thing to work on.
>
> Jonathan
>
> On 7/12/07, Roderic Page <r.page@bio.gla.ac.uk> wrote:
>>
>> I'm a little confused. Are you asking how do we find LSIDs that  
>> exist (i.e.,
>> some sort of discovery process)?
>>
>> I think most people intend LSIDs to be identifiers, e.g., they  
>> would appear
>> in a list of search results, be cited in a paper, web page, or other
>> document (such as an email message, a RDF document, etc.). In this  
>> way they
>> are like DOIs. I wouldn't  particularly want to know what DOIs are  
>> out
>> there, I'd want the DOI for  a paper I'm looking at, and given a  
>> DOI I'd
>> want a way to resolve it.
>>
>> There is a well described protocol for resolving a given LSID. The  
>> initial
>> step is looking up the authority server in the DNS. The FireFox  
>> client
>> doesn't seem to do this, by default it uses a CGI script to do
>> this look-up. Other clients, such as the Java, Perl, and PHP ones  
>> (at least)
>> talk to the DNS directly.
>>
>> Regards
>>
>> Rod
>

----------------------------------------
Professor Roderic D. M. Page
Editor, Systematic Biology
DEEB, IBLS
Graham Kerr Building
University of Glasgow
Glasgow G12 8QP
United Kingdom

Phone: +44 141 330 4778
Fax: +44 141 330 2792
email: r.page@bio.gla.ac.uk
web: http://taxonomy.zoology.gla.ac.uk/rod/rod.html
iChat: aim://rodpage1962
reprints: http://taxonomy.zoology.gla.ac.uk/rod/pubs.html

Subscribe to Systematic Biology through the Society of Systematic
Biologists Website: http://systematicbiology.org
Search for taxon names: http://darwin.zoology.gla.ac.uk/~rpage/portal/
Find out what we know about a species: http://ispecies.org
Rod's rants on phyloinformatics: http://iphylo.blogspot.com
Rod's rants on ants: http://semant.blogspot.com