[Geowanking] Better auto-discovery in the Geo-Web through "see” and “see also" links?

Sean Gillies sean.gillies at gmail.com
Tue Jul 28 01:01:27 PDT 2009


Perhaps use the "related" or "service" relation types from

   http://tools.ietf.org/html/draft-nottingham-http-link-header-06#section-6.2

If I were crawling for service capability documents, the inclusion of  
a type attribute for links would be most helpful. Is there an internet  
media type specified for OWS capabilities? I'd fetch and process these  
docs regardless of the link rel attribute.

Sean

On Jul 25, 2009, at 8:52 PM, Stefan Keller wrote:

> Yes, "web of services" sounds good - but connected to the general web.
> This is why I'm in favor of not only extending GetCapabilities but  
> also
>
>> 1. extending robots.txt (see [1])
>> 2. extending links/META tags in HTML ([1])
>
> So I mostly agree with you but are unsure why you don't comment on
> proposal 1. and 2.: you agree extending these three "protocols"?
>
> Put yourself into the situation of a so called focussed webcrawler:
> How do you guess which link to follow in the general web without these
> "see" and "see also" links?
>
> Yours, Stefan
>
> 2009/7/23 Sean Gillies <sean.gillies at gmail.com>:
>> Stefan,
>>
>> You're talking about capabilities documents linking to other  
>> capabilities
>> documents, forming a "web of services"? I think it's probably going  
>> to be
>> more profitable to exploit one of the bigger, more fertile webs  
>> (using HTML
>> or RDF), each of which have proven techniques for making your content
>> discoverable.
>>
>> Sean
>>
>> On Jul 23, 2009, at 1:17 AM, Stefan Keller wrote:
>>
>>> How about getting better auto-discovery in the Geo-Web?
>>> How about more special search engines for geospatial webservices?
>>>
>>> Approach: Step 1. a focussed crawler is referred to a webservice  
>>> given
>>> a (root) geo-website (“see” link) and step 2. the referrer and/or  
>>> the
>>> geo-webservice indicates one or more "friends" ("see also" link,  
>>> [2]).
>>> In other words, a porter points to the webservices - which is like a
>>> finding needle in the haystack - and he and the found needles point
>>> you to other needles!
>>>
>>> These are the proposals I've collected so far:
>>>
>>> 1. extending robots.txt (see [1])
>>> 2. extending links/META tags in HTML ([1])
>>> 3. extending (geo-)sitemaps (www.sitemaps.org), and
>>> 4. extending GetCapabilities
>>>
>>> What do you think?
>>>
>>> Regards, S.
>>>
>>> P.S. I favor 1,2 and 4. And I know I rather should ask geo-metadata
>>> specialists (which I'll do) - but I think this is the right place to
>>> start with.
>>>
>>> [1]
>>> http://www.openarchives.org/pipermail/oai-implementers/2006-November/001656.html
>>> .
>>> [2] http://www.openarchives.org/OAI/2.0/guidelines-friends.htm .
>>>
>>> _______________________________________________
>>> Geowanking mailing list
>>> Geowanking at geowanking.org
>>> http://geowanking.org/mailman/listinfo/geowanking_geowanking.org
>>
>> --
>> Sean Gillies
>> Software Engineer
>> Institute for the Study of the Ancient World
>> New York University
>>
>>
>> _______________________________________________
>> Geowanking mailing list
>> Geowanking at geowanking.org
>> http://geowanking.org/mailman/listinfo/geowanking_geowanking.org
>>

--
Sean Gillies
Software Engineer
Institute for the Study of the Ancient World
New York University





More information about the Geowanking mailing list