Re: Dynamic vs. Fixed World Views was Re: MARCXML to Topic Maps? MODS to Topic Maps?

Lars,

On 7/18/2010 8:39 AM, Lars Heuer wrote:
> Hi Patrick,
>
> [...]
>    
>> The light is better using topic maps but we had better go stand next to
>> potential users in the dark and understand where they are rather than
>> shouting at them to come towards the light.
>>      
> ;)
>
> Well, where is the money? :)
>
>    
That's a fair question.
> I understand what you're saying and I think we both like Topic Maps,
> but who should spend his time and energy to create the mapping with
> the thin hope to attract users?
>
> Well, the whole world should use Topic Maps, but.... where is the money?
>
>    

Hmmm, perhaps we should ask the NoSQL world where the money is that is 
driving their success?

It isn't possible to start off with a community that will pay you to 
build a product.

Imagine going to buy a car, computer, high-end sound system and being 
told, well, we are waiting for enough people to order one before we 
build any of them.

Curious what your reaction would be?

Still, your question is a fair one.

I think the answer lies in what payoff is there from using topic maps? 
In terms of actual ROI and/or benefits to users.

A couple of quantifiable examples:

1) Say a library system has a substantial ESL population. We know that 
even native English speakers can't effectively use Library of Congress 
subject headings. My suspicion is that an ESL population would have even 
greater difficulties.

Problem: Library in these days of budget cuts has no staff to create 
parallel subject headings in the language of the ESL population. Nor 
could they enter them in the library software due to a lack of 
multi-lingual support.

Solution: Create a topic map that is tied to the library records 
(possibly by the ISBN, more on that in a moment) and enlist members of 
the ESL community to create parallel subject heading understood by their 
population for the records. The catalog offers a "search in "language" 
options. Possibly even with unstructured prose descriptions of the text.

Results in greater service to the ESL community, library does not have 
to come up with funding for staff to create the subject headings, and, 
quite possibly a marketable product in the topic map since it can be 
used by other libraries with the same ESL community.

And libraries could collect numbers on how often that option was chosen 
for searching. Giving them the numbers to justify the investment in 
topic maps.

2) Legal discovery. Did you know that full text searching by lawyers 
retrieves only 20% of the relevant documents? That not just me, that has 
been demonstrated time and time again. With actual research. One of the 
issues is that subjects have different identifications.

But, another issue is that if I search the collection and discover some 
other identification, what do I do? Can I create a map of what I have 
found so that others will have their searches enhanced by my findings? 
Oh, yeah, that would take a topic map application wouldn't it?

Such an application would make searching of a document collection grow 
smarter (I am assuming good researchers, it could grow dumber) with each 
search of the collection.

I think quantifying the ROI in that case would be more difficult because 
of the size of document collections where it would be most useful but I 
suspect the user experience with a well-designed interface would be 
positive.

3) Medical research: Don't you hope the cure for what's wrong with you 
comes up in the first 20 hits in PubMed? Those are the records that get 
the most use in any search of PubMed. Despite the fact we know that 
differences in terminology are as prevalent in medical research as 
anywhere else.

Yes, the mapping would require expertise in medical terminology but 
indexing requires the same knowledge. Topic maps offer a way to engage 
the researchers, who are more likely to recognize parallel terminology 
than others.

Why? Well, for the same reason people who publish books want to get them 
reviewed. A desire for quality indexing and perhaps just a touch of 
vanity will encourage researchers to make sure mappings are correct and 
perhaps to contribute some on their own. Could even have a mappings 
contributed rating for researchers.

I am as guilty of this as anyone and that is to presume that technical 
excellence is enough. No, it is barely a starting point. We have to 
promote a technically excellent idea in ways that fit how people 
actually work. The SW has fallen into the same trap. Everyone "should 
be" this way. OK, but their not. Next question.

Note that third parties could generate topic maps to provide better 
access to legal collections, government regulations, public data sets, 
etc. All of which they could sell. Just a question of finding data that 
someone is interested enough in to pay money for.

I would think that a topic map of the US political contributions tied to 
actual legislation voted for by particular members of Congress would be 
a hot item. Would enable people to rate members of Congress on a ROC - 
Return on Contribution basis. Sure, you paid representative X $100,000 
and only got $10 Million in contracts. Had you paid representative Y 
$125,000 you might have gotten $30 million in contracts. Or do the same 
for lobbyists. ;-)

Or even less public but none the less useful applications. Such as the 
coming "uniform" medical record stuff in the US. Supposedly. Topic maps 
could provide a useful layer to mediate between whatever "uniform" 
system is required and current systems. Not a glamorous application but 
one that would fill a needed niche.

Hope you are having a great weekend!

Patrick




> Best regards,
> Lars
>    

-- 
Patrick Durusau
patrick-Q/[email protected]
Chair, V1 - US TAG to JTC 1/SC 34
Convener, JTC 1/SC 34/WG 3 (Topic Maps)
Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300
Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps)

Another Word For It (blog): http://tm.durusau.net
Homepage: http://www.durusau.net
Twitter: patrickDurusau