Re: Z39.50 on the web (and in print)

At 16:36 22-02-2002 +0000, Ashley Sanders wrote:

>Does not real Explain already have a F&N feature -- Surrogate
>Explain Databases as described by Denis Lynch in:

It does.

>All you're really looking for is an excuse to play with some new
>sexy XML tools. Who's bothered about boring old technology that
>is here now and works. Why _not_ invent a new variant of something
>we already have, thereby ensuring even less interoperability than
>before. Still it gives the programmers something new to play with
>and put on the CV.

Well.. in general I agree with you... in the case of Explain, I have to say 
that while I like the general sematic model, the choice of encoding the 
data structures in ASN.1 was seriously flawed...  they're cumbersome, 
unwieldy and unpleasant.

The second problem with Explain is that it is not only over-engineered... 
that would be OK. But at some point in the design of Z39.50 v3 someone 
decided to basically use it as a trashcan and throw *EVERYTHING* in there.. 
preferably twice for good measure if there was any doubt. By the time I 
began going to the ZIG meetings, the phrase "oh, we'll just put it in 
Explain" was still a standing joke amongst the old-timers. I respect the 
brainwork behind Z39.50 tremendously... I've dedicated a large portion of 
my professional life to working with it... but Explain is very clearly the 
result of some terrible mass psychosis. It is broken.

I personally started to get tired of advocating it when, after several 
years, I failed to find two people who could agree on something as dead 
simple as how you describe which searchable attributes you have in your 
database.

> > Someone define the exchange/negotiation mechanism and data format and I'll
> > race ya to implement it!   :-)
>
>Can I suggest z39.50 and Explain records :-)

I've done that.. not sure if we still have the server running, but I'll go 
check....

But realistically, for this kind of thing to work, it has to be technically 
dead simple. A full Explain database is not really simple by any stretch of 
the imagination.. neither to build or to work against.

As much as I hate to admit it, I think Explain has had its chance... it is 
going to take a lot more than provocative posts on this list to change that.

Now, if someone were to propose profiling (read: slashing) Explain down to 
the essential information categories and recast them in a sensible syntax 
like GRS-1 or XML, then I might start to feel better about things.

--Sebastian
--
Sebastian Hammer, Index Data <http://www.indexdata.dk/>
Ph: +45 3341 0100, Fax: +45 3341 0101

Received on Friday, 22 February 2002 16:59:44 UTC