In article <1globmINNn4t@crcnis1.unl.edu>, price@helios.unl.edu (Chad Price) writes:
> clinton@acs.ucalgary.ca writes:
>
>>The Alpine Club of Canada is looking into "buying into" a database
>>developed for the Seattle Mountaineers in Advanced Revelation.
>>Naturally, our needs are not exactly the same, but from the talks we
>>have had with the person who is actually using the database, it is very
>>close to our needsd: Basic membership tracking, event scheduling, hut
>>bookings, donor tracking, etc. But we will have to make modifications.
>
>>Here's the questions...
>
>> Relational
>> How close does it get to Codd's 12 rules?
>
> Fairly, but some rules are intentionally broken: ARev uses "multi-valued"
> fields. In other words, a single row/column intersection could contain multiple
> values rather than a single value as in a completely normalized database. This
> is functionally and performance-wise, a strong advantage. It is fairly typical
> for a primary table to have one or more fields which are 1:many relationaships.
> ARev allows this 1:many to hold true, avoids deletion and insertion anomalies,> and allows the higher performance of a single table data entry screen or report
> program.
I agree... Codd's rules are important to providing a solid RDBMS platform, but
don't bypass products because they break some of the rules... and Chad's
example here is one very powerful feature that ARev supports...
>> Does it support integrity rules in the dictionary?
> Not directly, but you can create "symbolic" fields which are in effect programs
> to ensure data integrity, and there are hooks into the data entry screens to
> allow inclusion of these symbolic fields for data validation. In fact, if you
> write the validation code before you create the screen, it is automatically
> included.
Hmmmm.... actually, I would say ARev _does_ support integrity rules in the
dictionary... ARev data dictionaries contain 'Edit' constraints which can be
any of a wealth of 'canned' constraints [dates, numeric fields, pattern
matching, etc..] and can also contain custom constraints through "symbolics"...
once you define these edits, they are bound to the data element in any future
reference to that element - but this is limited to Paint... so a little caution
is prudent - you CAN bypass the edits...
>> Query, Forms, Reports
>
> .......It has its own 4GL built-in
[called R/List]...
>> Does it have an interactive command line for ad-hoc queries?
> Yes, and you can store the ad-hoc queries and add them to a menu
>
>> How easy is it to build a form?
> Simple forms: trivial - there is a "Quick Paint"
> Complicated forms: up to you - the entire power of the built-in R/Basic
> programming language is available.
>
>> Programming
>> How 4GL is it?
> Which part.. Some is (TCL, SQL, R/List query language), and some isn't: the
> R/Basic programming language.
[I've pulled a few pieces of the original post together here]...
.... and one of the [IMHO] most powerful features available to the programmer
is the ability to build truly custom routines from R/List... R/List is a very
powerful query/report generator, and can be used in ad hoc or 'canned' modes..
and in those situations where it just can't quite do what you want, there is an
option to generate R/Basic source code - from the R/List query - and then
modify the source, compile, and you are off and running... it's wonderful - you
can build [and test interactively] all of the database access logic in an
R/List query, and then customize the source code for unique requirements... and
once you have it running, the new routine can be added to the vocabulary of the
end system as an extension...
>> Documentation
>> How does it rate?
>
> I can't really tell. To me its quite good, but after 7-8 years of reading it, I
> understand their philosophy and I know where to look for things.
...sigh... this is one of the areas where I've often run into problems... the
ARev docs are much better than G2B... but it still takes some time for folx who
are not familiar with the product... recognize that ARev [and G2B] are
extremely powerful and adaptable products... there is so much to know about
these platforms that it is difficult to present _everything_ in a simplistic
manner. I find what I am looking for fairly quickly, but then again, I know
*where* to look... this is not always the case for beginners [and I've watched
four people go through the learning process recently.. initially there is a lot
of frustration with the docs, but after some exposure they do much better]
>> If you can
>> make comparisons to Oracle on one end and R:base on the other...
hmmmm... *WARNING*... these are the ravings of a confirmed Revelation
fanatic... [grin]... I started using Rev products [1985-1986] because no other
product could support our requirements... and R:base was one of those... I
haven't seen it recently, so it wouldn't be fair to comment... but I've worked
with Oracle [too recently]... and quite frankly I am not impressed... the
product is cumbersome, resource intensive [you'll need a LOT of RAM and HD
space] and very expensive...
>>Is it strictly DOS?
> No - I run both the DOS and Windoze versions under OS/2, which is what I run
> both at work and at home. I'm very happy with it in the multi-tasking
> environment.
... and the last time I talked with the folx at Hunter... their port to UNIX
was in process and should be avail sometime spring 93... oh boy oh boy!
>>What does a license cost?
> Too much! list is >$1000, but it is available for less. I paid $250 as an
> upgrade from G2B, some places like Telemart sell it for around $700. Run-times
> are between $150 and $250, depending on where you get them. 5-user LAN bump
> disks are around $495 (list) minus whatever discount you can find.
Yup... Telemart advertises the single-user version for $707... and I just
ordered a server version for $606 and a 3-user add pack for $477... their
