Re: Can we solve this -- NFNF and non-1NF at Loggerheads

"David Cressey" <david.cressey_at_earthlink.net> wrote in message news:wg3Pd.16699$wK.469_at_newsread3.news.atl.earthlink.net...
>
> <lauri.pietarinen_at_atbusiness.com> wrote in message
> news:1108076484.570592.182420_at_o13g2000cwo.googlegroups.com...
>
>> The application and the database communicate "only" thru SQL, so
>> there is a lot of metadata that is not communicated between the
>> layers, at least not without a lot of effort. I have to confess
>> that I don't have a lot of experience with 4GL's except for
>> SAS (it had a lousy DB, though).
>>
>> What I mean is that the app environment is not aware of
>> the structure of the DB without lot's of (hand-) coding.
>>
>
> Perhaps that's why I look back on Datatrieve so fondly. While I would
> hardly call Datatrieve a 4GL by any stretch of the imagination, it was
> possible to build a small information system in it with very little hand
> coding of common activities.
>
> It contained its own primitives for restrict and join. Project was only
> added much later. The Datatrieve "programmer" did NOT have to learn SQL.
> In fact, Datatrieve didn't even use an SQL like interface until the late
> 1980s.
> In short, the power to hassle ratio of Datatrieve was quite favorable.
>
> The world has turned many times since those days. I would NOT want to
> "bring back Datatrieve". But something that it captured has NOT been
> captured by the later, more ubiquitous products. I'm not exactly sure
> waht
> that was.

I'm with you on that. While I was one of the folks in IT who struggled with satellite systems springing up in datatrieve and housing data out of synch with centralized databases, I also saw departments who were happy with the product -- yes, happy! They could do what they wanted to do. I'm searching for the aspects of those old systems that really brought a high delight-to-cost ratio and see if we can get those in the systems moving forward. SQL seems to be a big factor in losing ground on this front continuously over the years. The data model that prompted SQL (yes, yes, I know relational theorists, in general, don't like SQL either) also seems to have done some harm in this regard.

Cheers! --dawn Received on Fri Feb 11 2005 - 16:19:50 CET