[ZOOM] Re: Scope of ZOOM (was Hello etc)

Sebastian Hammer quinn at indexdata.dk
Wed Nov 7 14:57:44 CET 2001


At 12:41 PM 11/7/2001 +0000, Robert Sanderson wrote:

>I have the feeling that you're missing the same point that the SRW (ex
>ZNG) crowd are - by the time you've added everything that 'Z Classic' does
>back into the new ZiNG umbrella, it's going to be just as complex and
>just as unmanageable and it's easier for people to write their own code
>(on top of some libraries obviously) than to learn how someone else has
>done it.

I don't think there's an easy answer here.

In part, Z39.50 is complex *because* it has been relatively popular, in its 
own obscure way. It's a accumulated a host odd features over the year 
because someone has come by and said "wouldn't it be nice if...?". But this 
is not unique to Z39.50... just look to the collected specs for HTML, HTTP 
& co. for even grimmer examples. In fact, I would guess that Z39.50 may 
actually be cleaner in its design than a lot of protocols out there exactly 
because the design has evolved in a fairly consensus-based process. It's 
probably also slowed down its evolution, but that's another matter. But I 
happen to believe that with some horrenodus exceptions, a surprisingly 
large proportion of the protocol is basically sound.

A big reason why ZOOM (and SRW?) will be neater than Z39.50 is because we 
can design things into a coherent whole and straighten out the bulges that 
were introduced into the protocol to support backwards compatibility. That 
doesn't mean that we have to include everything from the start, but it does 
pose the challenge that we should have a framework that can be extended 
with new features as they become relevant, without breaking down.

I fully expect that the ZOOM specification will evolve over time. If it 
becomes more popular, people will ask for more features. For me, the 
primary thing that we gain from an OO-style of interface is really named 
parameters that carry implicit default values -- because this allows us to 
include fancy features without bothering the poor saps who just want a 
dead-simple client.

We can sit here and decide that ZOOM will only be an introductory toy for 
first-timers who need to write dead-simple applications.. I've even felt 
that was the best way to go on occasion.. but it just won't work, because 
people WILL want to write fancy, richly featured apps with ZOOM... and we 
will not turn them away, because Some of Them Will be Us.

--Sebastian




More information about the ZOOM mailing list