Re: Ann: COLD - Coloring the Data Web

Hi Hugh,

Now sameAs is added :-) Congratulations!
http://cold.aksw.org/index.php?page=tools

On 04/03/2013 12:15 PM, Hugh Glaser wrote:
> Yes, I think that works nicely:
> http://www.sameas.org/?uri=http://aksw.org/MichaelMartin
Yes that looks quite good. I like my color.

We had some trials using SHA256 / SHA384 / SHA512 in comparison to MD5 
as basis for hashing URIs.
But as a result MD5 creates the most beautiful colors :-)
> And of course less load on your server and the network.
>
> Hmmm, yes, I did think the 24 bits would be a limitation, but moving to 64 or 128 bit colours would solve the problem.
> All people would have to do would be remember the 128 bits, which is of course pretty easy.
> Or we could have a bitty-style service for the colours.
> Or even a service that translated symbols into colours, based on domains.
> Oh, hang on, that's what your service does.
of course!
>
> Tee hee!
>
> On 3 Apr 2013, at 10:22, Michael Martin <martin@informatik.uni-leipzig.de> wrote:
>
>> Hi Hugh,
>>
>>
>> On 04/02/2013 01:07 PM, Hugh Glaser wrote:
>>> Hi Martin,
>>> Brilliant initiative!
>>> So many steps forward for Linked Data over this weekend.
>>> I have hacked sameAs.org to conform to what will clearly become a crucial standard.
>>> However, this has exposed what may be a shortcoming.
>>> Your colours (sic) are related to the IR; but would it not be much better if they related to the NIR?
>>>
>> This is an interesting point, but seems to be more complex :-)
>> But we can offer that as an further output, so users can decide what to use themself.
>>
>> As a starting point it would be perfect if http://dbpedia.org/resource/Edinburgh is being colored the same way in different tools.
>> Maybe this could lead to a more frequent re-usage of URIs. A further advantage is that it can be reimplemented without using any web service.
>>
>> A second step could then be if http://dbpedia.org/resource/Edinburgh and all the other 120 listed resources about the same thing are coloured the same.
>> The usability of tools would be heavily improved but needs a further pre-processing step.
>>> You will see that www.sameas.org/?uri=http://dbpedia.org/resource/Edinburgh
>>>   illustrates the point.
>>> (and takes rather a while to load, while it finds all the colours and renders the stuff!)
>>> Would it not be better if COLD was "sameAs" aware?
>>> You could query sameAs.org and find the colour of the canonical URI (the foaf:primaryTopic in the RDF), and render that colour for any of the URIs?
>>>
>> We can offer that approach as well. and integrate respective code snippets on our snippets page ( http://cold.aksw.org/index.php?page=snippets).
>> I will add sameAs.org on our tools ( http://cold.aksw.org/index.php?page=tools) page? Is that ok for you? :-)
>>> Or maybe offer it as an option.
>>> I think I will go to only showing the colour for the main URI soon, as the intention is clearly to establish colour association in the user's mind.
>>>
>>> One day we will be able to dispense with URIs altogether and just use colours!
>>>
>> ok, there is a 24bit limitation but i would like that idea as well :-)
>>
>> Best
>> Michael
>>> On 2 Apr 2013, at 10:34, Michael Martin
>>> <martin@informatik.uni-leipzig.de>
>>>
>>>   wrote:
>>>
>>>
>>>> Hi Ben,
>>>>
>>>> On 04/01/2013 10:38 PM, Ben Companjen wrote:
>>>>
>>>>> Hi Michael,
>>>>>
>>>>> There is a small bug in the Turtle representation: in e.g. my colour
>>>>>
>>>>>      
>>>>> <http://companjen.name/id/BC>
>>>>>    a   rdf:Resource ;
>>>>>          cold:colour loc:26bd27 .
>>>>>
>>>>>      cold:color a owl:AnnotationProperty ;
>>>>>          rdfs:label "color"@en ;
>>>>>          rdfs:domain rdf:Resource ;
>>>>>          rdfs:range dbpo:Colour .
>>>>>
>>>>> ... you mixed up "color" and "colour". That's all :)
>>>>>
>>>> thank you for reporting this small bug. Now it should be fixed.
>>>>
>>>>> By the way, I really like my colour!
>>>>>
>>>> Yes your color is quite good.
>>>> Its even more fancy than my one:
>>>>
>>>> http://cold.aksw.org/index.php?iri=http%3A%2F%2Faksw.org%2FMichaelMartin
>>>>
>>>>
>>>> Maybe we should open a contest about beauty WebId's :-)
>>>>
>>>> All the best
>>>> Michael
>>>>
>>>>
>>>>> Regards,
>>>>>
>>>>> Ben Companjen
>>>>>
>>>>> On 1-4-2013 20:37, Michael Martin wrote:
>>>>>
>>>>>> Dear all,
>>>>>>
>>>>>> On behalf of AKSW research group [1] I'm proud to announce an innovative
>>>>>> approach for coloring the Data Web. The monochromacity of the Data Web
>>>>>> is widely perceived to be the main obstacle for a wider deployment and
>>>>>> penetration of Linked Data and Semantic Technology (cf. e.g. [2]).
>>>>>>
>>>>>> So far, no unified algorithm existed for coloring the Data Web. With
>>>>>>
>>>>>> http://cold.aksw.org
>>>>>>    we developed a key base technology able to color
>>>>>> URIs and IRIs (future work will focus on literals, whole triples,
>>>>>> containers, reifications etc.). Features of COLD include:
>>>>>>
>>>>>> * globally unique URI/IRI coloring algorithm
>>>>>> * cross-application color consistency
>>>>>> * ensuring color fidelity
>>>>>> * built in color attack prevention
>>>>>> * support for vocabulary/ontology coloring
>>>>>> * 24bit / 16,777,216 color support
>>>>>> * integrated RGB support, extensibility for other color models
>>>>>> * example implementations in five programming languages
>>>>>> * small memory and code footprint
>>>>>>
>>>>>> We deem COLD to be the key technology for the ultimate breakthrough of
>>>>>> semantic technologies. COLD is already implemented in a number of tools
>>>>>> including CubeViz [3]. Please beware of brand infringement, due to color
>>>>>> trademark protection (cf. [4]).
>>>>>>
>>>>>> Best,
>>>>>>
>>>>>> Michael Martin
>>>>>>
>>>>>> [1]
>>>>>> http://aksw.org
>>>>>>   
>>>>>> [2]
>>>>>> http://purl.org/colors
>>>>>>   
>>>>>> [3]
>>>>>> http://aksw.org/Projects/CubeViz
>>>>>>   
>>>>>> [4]
>>>>>> http://brandcolors.net/
>>>>>>   
>>>>>>
>>>>>>
>>>> -- 
>>>> Michael Martin, M.Sc.
>>>> Department of Computer Science, University of Leipzig
>>>> Research Group:
>>>>
>>>> http://aksw.org/
>>>>
>>>>
>>>> Homepage:
>>>>
>>>> http://bis.informatik.uni-leipzig.de/MichaelMartin
>>>>
>>>>
>>>> Phone: +49 341 97-32322
>>>>
>>>>
>>
>> -- 
>> Michael Martin, M.Sc.
>> Department of Computer Science, University of Leipzig
>> Research Group:
>> http://aksw.org/
>>
>> Homepage:
>> http://bis.informatik.uni-leipzig.de/MichaelMartin
>>
>> Phone: +49 341 97-32322
>>
>


-- 
Michael Martin, M.Sc.
Department of Computer Science, University of Leipzig
Research Group: http://aksw.org/
Homepage: http://bis.informatik.uni-leipzig.de/MichaelMartin
Phone: +49 341 97-32322