Popular links

Topic maps


Claudio Monteverdi

Alexander Johannesen

Technologies used
topic map logo
xSiteable logo
Mon, 3 May 2004 13:00:00 GMT

Notice! This blog is no longer updated as such, and the new spot to point your feedreaders and blurry eyes are https://shelter.nu/blog/

This also means no more comments here, and especially not you spammers, you filthy floatsam of the internet!

An essay about Topic Maps : The answer to everything. I think. Maybe.

I've just published an essay about Topic Maps, the concepts and how it all works. It includes name calling, Amazon.com bashing, conceptual waffling and philosophical musing of RDBMS as compared to "do the right thing." Enjoy.

Permalink (Mon, 3 May 2004 13:00:00 GMT)| Comments (18) | Knowledge and information Topic maps
Mark Siegal ( Tue, May 4 2004 )
Nice essay. (Got here via http://www.informationdesign.org/archives/002442.php#002442.) I'm intrigued by the rest of your writing and by xSiteable.
By the way, how close are you to xSiteable 0.9? I would love to muck around with it. This would be an excuse to finally learn PHP.
Mark ( Tue, May 4 2004 )
xSiteable 0.9 is close ... ish. I'm a bit on the slow side of it, being a spare-time project when I've got none, but ... sooon. :)
BTW, all PHP has been taken out of it, so I'm sorry to say you won't be needing it, unless you feel called and want to take over the PHP part I took out.
Mark Siegal ( Wed, May 5 2004 )
That's funny... now I'll just have to learn PHP for some other open source project. What's the administration package written in now?
Alex ( Wed, May 5 2004 )
Oh, it's worse than that; I've taken the whole admin package out. I can't support it as I have hardly time to do the bare basics. :(
Having said that, the 0.95 version is modulised in such a way that it would be easy to put it into a caching framework and have the whole shebang as a server-side CMS. I'll make some sketches.
( Sun, May 9 2004 )
I'll print out the essay at work tomorrow, and will read it on the subway (as all other litterature I print out these days) to and from work. Looking forward to it.
Would you be interested in having me port xSiteable to ASP.NET, to make it more vendor independant and cross-platform? I haven't tried the PHP version yet, so I have no clue on what it offers, but I've been thinking a long time about writing a CMS/Weblog-ish thing with support for alternate representations of the content, like XHTML, RSS, XTM, RDF and Atom. I've also thought about writing a Windows application as an alternative to web for writing articles, and to use the Atom API as a standard posting interface.
Can these ideas be combined with what xSiteable is today? I'm sure there would need to be some tweaks and modifications, but I really need a kick in the ass to get myself going on this, and xSiteable seems like a great starting point as well as it would be great to have someone on the «team» that knows a little more about open source develoment (especially CVS) than myself. So, whaddayasay? :-)
PS: Would it be too much to ask to have your comment form updated a little? What I would like is a 'Name' field, a 'URI' field (either e-mail or homepage address), a 'Remember me' checbox, and a [Preview] button. :-)
Alex ( Tue, May 11 2004 )
Hi Asbjoern,
Funny how my essay on Topic Maps has turned here into a discussion on mostly xSiteable. :) Must be something up with that.
Anyways, xSiteable's PHP tool is per next version unsupported, as I don't have the time. Porting that thing into a) something more maintainable and b) something that support version 0.9x can only be seen as a good thing. But if you really want to do it, let me know. My plan (and the biggest reson for the 0.9x rewrite) is to modularise the whole thing, make it less dependandt on platforms, ontologies and specific technologies. In fact an ASP.NET version sounds pretty cool, and if you do it right, it can be further ported easily to others, inclusing back to PHP if someone wishes.
Anyways, give it your best shot. The 0.9x rewrite has made the system a lot more appealing to anyone wanting to write a simple cache and editor (CMS) around it. let me know at alex at shelter dot nu.
Oh, and about the comments boxes and such on my site; it is a homebrew system, and as nice as 'remember me' and previews and such might be, I'm not sure I've got the time right now. :) I'll chalk it up on the list of things to do when things around here settle down. Ha!
Alex ( Tue, May 11 2004 )
Oh, and if anyone has any comments on the actual essay, PLEASE chalk them down here. I've had thousands of reads of the article, but only two or three comments. I need some feedback! Please!
matti ( Tue, May 11 2004 )
Well written essay. I have just started getting into topic maps, like studying TAO and other 'canonical' texts and your essay inspired me even more. I liked your comparision between relational databases and topic maps, because I think that most of us is familiar with RDBMS. The notice 'The truth about relational databases is that they really are Topic Maps that are trying to get out' is brilliant. Now I'm eagerly waiting for the following part :)
Tom Passin/ ( Thu, May 13 2004 )
This gives me an opportunity to share my view of the relationship between topic maps and relational databases.
Consider a table in a well-normalized relational database. Suppose its primary key is atomic rather than compound. Then the primary key corresponds to a topic. That is easy to see because all data in the row is 'about' the primary key, which must not depend on any of those pieces of data.
Each column in the table, aside from the primary key, contains either a data value or a reference to another table (i.e., a foreign key). Cells with data values correspond to occurrences, and the column label corresponds to the name of the occurrence type.
A cell that contain foreign keys corresponds to an association between the primary key topic and the foreign key topic. The column name corresponds to the role, and the label of the relationship between the two tables corresponds to the name of the association type.
Simple, eh?
What about compound keys, such as you would have for a join table? They clearly correspond to associations, since the individual parts of the compound key are (usually) foreign keys. These associations could be three-way or more, depending on how many parts the compound key has.
If a table with a compound key has data value cells as well, this corresponds to reifying the association and adding the data value cells as occurrences to the reifying topic.
What about subject indicators and scopes? Well, they do not really have counterparts in a relational database.
What about poorly normalized relational databases? Here you are on your own, but a good plan would be to normalize it decently first, or at least to figure out a way to transform it as if it had been normalized when converting to a topic map.
Alex ( Thu, May 13 2004 )
Tom, that is quite interesting, and I think I remember you going on about this also with your bookmark application for TM4Jscript? I've had similar notes (and support in xSiteable) for pure CVS exports from Excel.
What would be really good - and a lever for RDBMS people to use TM - to create a tool that questions a DB, receives the setup, and creates an XTM (with or without the actual data) of it. How easy it would be to convert the gentiles!
Any takers? Dang, there just is too much interesting stuff floating around these days in terms of Topic Maps.
Tom Passin/ ( Sat, May 15 2004 )
I have done a partial mapping by hand of one moderately complex database. I used Cold Fusion to create the XTM, IIRC (I have Cold Fusion at work). It was not too hard - I did most of it in an afternood, and I was learning as I went - and I could see getting it to happen more-or-less automatically, but only because that database is fairly well normalized.
I don't see how it could be done completely for the general case, because too much of the semantics of the database design are hidden in the names and data structures (not to mention triggers and stored procedures). This includes any denormalization. A human has to understand the semantics of the database for a map to be produced sucessfully, IMHO.
But given a database in perfect 3rd normal form, you could probably do pretty well automatically, most likely not much tuning would have to be done by hand.
Tom Passin/ ( Sat, May 15 2004 )
I just found the RDBMS - topic map work I mentioned above. I had forgotten that I used Cold Fusion to create an intermediate xml format, and then used an xslt stylesheet to transform to xtm. It was pretty easy to write the CF page.
I created the basic ontology from the table and foreign key entity names by hand (plus my knowledge of the design and semantics of the database), then ran some sql queries and let CF populate the rest of the file with actual data from the queries.
Alex ( Sun, May 16 2004 )
Tom : That sounds like another OSS package that could be handy to create. :) Support for all sorts and not just CF would be good too. Say integration against PHP PEAR RDBMS module woul make sure you could hook up to quite a lot of DB's.
( Sat, Apr 22 2006 - 04:04:55 )
[URL=http://www.google.com/sf905hf]gi5geg[/URL] be$t http://www.google.com/sf905hf gi5geg brin brin
[URL=http://www.google.com/349gfff]45hgfhg[/URL] be$t http://www.google.com/349gfff 45hgfhg be$t rulez
rulez http://www.google.com/345gfrw dshf95g brin [URL=http://www.google.com/345gfrw]dshf95g[/URL] rockz
( Sun, Apr 23 2006 - 01:04:20 )
O, how this spring of love resembleth
The uncertain glory of an April day!
( Sun, Apr 23 2006 - 02:04:46 )
If we had no winter, the spring would not be so pleasant; if we did not sometimes taste of adversity, prosperity would not be so welcome.
( Sun, Apr 23 2006 - 03:04:31 )
April comes like an idiot, babbling and strewing flowers.
( Sun, Apr 23 2006 - 03:04:51 )
O, how this spring of love resembleth
The uncertain glory of an April day!