NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/database/7948

Internet Message Format | 1992-11-19 | 4.0 KB

Path: sparky!uunet!gumby!wupost!crcnis1.unl.edu!price
From: [e-mail] (Chad  Price)
Newsgroups: comp.databases
Subject: Re: DB Comparison - Revelation/ARev comments
Date: 19 Nov 1992 16:52:00 GMT
Organization: University of Nebraska--Lincoln
Lines: 67
Message-ID: <[e-mail]>
References: <[e-mail]>
NNTP-Posting-Host: helios.unl.edu

[e-mail] writes:

>Revelation - I worked with Rev/G for about 2 years.  It drove me right up a
>      wall.  I managed to hack it into something reasonable, but I denounce
>      what IMHO is an insurmountable design flaw - every field is variable
>      length.  Rev fans no doubt are about to flame me, but consider: in most
>      database applications, 90% of your data is of known *fixed* length or at
>      ignores this little fact of life with no profit.  I tried ARev v1 and
>      found it overwhelmingly cumbersome.  Given the wide variety of reasonably
>      good products on the market, I can't see investing the time in Revelation
>      (zip up my Nomex suit for the flames-to-come).

Minor flame on:

    Rev G was great compared to the dBase II & III that were available at
the time! Problems mostly revolved around documentation which was so incomplete
as to not tell you enough to get the job done. Once you found the undocumented
system subroutines and a way around the very clumsey screen painter, you were
OK. Note that the painter, not the screen run-time environment was the problem.

I agree re ARev 1.0 and 1.1. They were/should have been beta releases of what
2.0 and 2.1 became. 2.1 to 2.12 are really extremely nice development
environments. Lots of flexibility to build whatever you need. There are still
some minor problems with undocumented side-effects (see my postings in the last
couple of weeks about global variables being tromped on).

Version 3.0 is another change in paradigm in some ways. They have changed
enough of the internals that its taking a while to get used to the changes.
Some are clearly for the better: ANSI SQL compliance for all table and field
names, far better use of the mouse with the screen painter, some additional
convenience features... Drawbacks include a 2.1 to 3.0 conversion program which
was released with far too many bugs that require manual workarounds! There seem
to have been some arbitrary changes in table names and consolidations of system
tables (or at least the name change reasons are not stated and are not obvious
to me).

As to the variable length fields, if that's a problem, use the dBase
Environmental bond and use native dBase fixed length field formats. The
capability to handcuff yourself with fixed length fields is built in! (but why
use it?). :-)

As much as the company's upgrade cost and run-time costs have made me search
for a new database to do my work  in, I keep wandering back to ARev because I
cannot find a replacement with as much easy-to-use power to replace it! FoxPro
doesn't cut it - too much programming to build stuff that ARev has built-in and
the Fox documentation is abysmal! Most of the xBase systems are drastically
handicapped by the database model forced on them by a non-relatoinal language.
Straight SQL products are handicapped by the lack of real-world procedural
extensions and by the forced fixed length fields (ever try to write a
bibliographic database to hold abstracts? Memo fields are a step in the right
direction, but incomplete).

Also the Rev Tech attitude towards paying for technical support is also bad
news for the small developer who can't aford to shell out a lot of money to
buy support for something he only uses evenings and weekends. (I got lucky with
my current job, and use it at work too, but we still can't justify the amount
of $$ that Rev Tech wants for support beyond the initial 3 (?) months of free
telephone support.

Final Conclusion: An excellent product from a company whose attitude towards
small developers stinks!!

(Flame (?) off)
--
chad
[e-mail]
[e-mail]