OK, now I think I also got the point. May I summarize?
Basically you (glenn) suggest to make typed links and attributes the
same thing. You also suggest to change their syntax, but that's not the
main point of your suggestion, so I won't dwell on that. By the way, I
personally prefer the term "typed link" for semantic relation, because I
think it's the closest to actual wikipedia word usage (and it's what we
suggested at the Wikimania, where it was met with quite some approval).
I guess the idea is, that it's easier for a non-geek to use. This could
be. This is something that needs to be investigated by HCI people.
What's really interesting in your suggestion - because it isn't possible
with our approach yet - is this:
>> The average height of a Martian female is [[:=2.9m|2.9 meters]].
>> Modern sea kayaks can weigh as much as [[:=65lbs]] and as little
>>as [[:=40lbs]].
This is to annotate all data values with units through the text, not
just the ones that are attributes. This would be a new possibility.
But - what I don't get is, what for? If it is not an attribute, meaning,
if you don't connect the data value with the concept of the article, but
just have a value (possibly with an unit), what can you do with it? Find
all occurences of 2.9 meters in the Wikipedia?
So, I think I see the added new feature of your suggestion, but I seem
to miss the added value. I'm sure there is some, but I am missing it,
sorry for that - please, can you elaborate on this? What's the usage
scenario of this feature?
Well, I do see one: to have the possibility to add highlighting and unit
supoort. So one could on the Martian female page move her mouse cursor
over the height, and get the conversion to feet. In this case, the
proposed syntax would make sense and could be included easily I guess,
but it wouldn't really make a difference wrt the semantic backend.
I hope I understood the suggestion right.
Best,
denny
Markus Krötzsch wrote:
> On Wednesday 07 December 2005 04:32, glenn mcdonald wrote:
>
>>OK, I'll try again and see if I can say this more simply.
>>
>>
>>Currently you are adding two new features to wiki editing:
>>
>>
>>Feature A1 (currently known as Semantic Relations) allows the
>>addition of semantic information to a link, for example:
>>
>> [[has capital::Berlin (Germany)|Berlin]]
>>
>>
>>Feature A2 (currently known as Semantic Attributes) allows the
>>addition of semantic information, and optionally units of
>>measurement, to a non-link data value, for example:
>>
>> [[length:=45000km|45,000km]]
>>
>>
>>It seems to me that the features you are adding would be more
>>powerful if they were conceived slightly differently:
>>
>>
>>Feature B1 would isolate the insertion of quantities with associated
>>units of measurement, which it seems to me is a generally useful idea
>>that should not be restricted to use in only semantic contexts. For
>>example:
>>
>> The average height of a Martian female is [[:=2.9m|2.9 meters]].
>> Modern sea kayaks can weigh as much as [[:=65lbs]] and as little
>>as [[:=40lbs]].
>>
>>(I'm not sure [[:=...|...]] is the best syntax, but it might be OK.)
>>I might call this feature "Unit Values", so we would say that the
>>above sentences contain Unit Value of 2.9 meters, 65 pounds and 40
>>pounds, with or without optional display strings.
>>
>>
>>Feature B2 would combine the *semantic* parts of both A1 and A2 to
>>allow the addition of semantic information to any kind of data value,
>>whether it is a link, a numeric value with units, or just some text.
>>For example:
>>
>> [[has capital :: [[Berlin (Germany)|Berlin]] ]]
>> [[is capital of :: [[Germany]] ]]
>> [[length :: [[:=45000km|45,000km]] ]]
>> [[nickname :: Snuffy]]
>>
>>I thought of calling this feature "Data Properties", so we would say
>>that Berlin has the Data Property that it is the capital of Germany.
>>My wife's less database-y idea was to call it "Fact Tags", so we
>>would say that Berlin has a Fact Tag stating that it is the capital
>>of Germany.
>>
>>
>>Is that clearer?
>
>
> Yes, definitely! (though my head is still somewhat jetlagged -- I just arrived
> in San Francisco) I understand your concept and why it has some advantages
> over our approach. But it also has some disadvantages (especially: more
> complicated syntax). I would like to combine the advantages of boh ideas, but
> cannot come up with a nice solution right now. I will have to sleep over it
> (and maybe discuss with some colleagues ...
>
> Anyway, we will thoroughly consider these things (probably I will not find
> much time during my current trip, so it might be Tuesday before I have more
> results). But feel free to make further suggestions in the meantime (i will
> read mails, but do not know if I have time to answer).
>
> Best regards,
>
> Markus
>
>
>>glenn
>
>
--
Denny Vrandecic
Institute AIFB, University of Karlsruhe (TH)
http://www.aifb.uni-karlsruhe.de/WBS/
Blog: http://semantic.nodix.net
|