Would anyone care to contribute relevant feedback (positive or negative) regarding the use of variants of the Pick database (other than AREV) which are still being marketed to develop robust, multi-user software applications?

Does anyone have specific experience with D3 from PICK systems?

How do the 'other' multivalue databases stack up to AREV/OI? Are there any out there worth considering?

I am a long time AREV developer and absolute AREV fan, about to embark on a significant new development effort intending to use AREV as the back end, but want to know from the user community if there are any other 'similar' products out there worth serious consideration.

I have enormous respect for RTI and for Revelation software products in general, but want to ensure that I am not overlooking viable alternatives.

Please feel free to respond via this forum, or to gsorbo@ziplink.net

Thanks.

Gene

Gene

Your best bet would be to go to the Spectrum show which is the Pick industry trade show as all the Pick based (with the exception of RevSoft) vendors attend.

I appear to have tossed out my current issue of Spectrum magazine so I don't have the contact information handy (handy magazine as they've had a long running series of articles on using MV in Visual Basic applications through ODBC etc.).

However you can look at the following websites:

UniData, UniVerse & Informix

Pick Systems

Most of the old Cosmos team's latest There may be a link to Spectrum on any of the above. I don't know for sure. and of course [url=http://www.livewater.com]Livewater

Warren,

Thanks for the input. Visualworld.com looks especially interesting.

Gene

Gene

My far and away favourite multivalue system was/is Universe Pick. Combined with Universe Objects you could probably make a very good GUI interface to a multivalue database. I haven't used Universe Objects but I hear it is similar to VB. It was also possible to use Universe with ODBC (despite the multivalues) so you could load the data up into MS Access or whatever else tool your users liked for adhoc reporting although the TCL was very good.

The tool overall is way more powerful than AREV and OpenInsight combined. Ie everything you can do in these two you can do in UniVerse and more.

UniVerse runs happily on either Unix or NT. It is not prone to corruption when the network crashes. It does not have limits to variable sizes (other than perhaps available hard drive space). It does not need special tools to make it go fast on a network. It is very very fast compared to OpenInsight.

The locking works, you don't have to set special flags on your forms to make the locking work.

D3 - I don't know anything about the D3 end but the PICK Systems end ie their TCL and Editor were horrible. The TCL was missing a lot of useful commands. Editor was a line editor, which you couldn't split lines or join them or copy them or move them or delete a block of them or anything that makes writing large programs useful.

I would choose Universe over D3 any day.

Mcdonnell Douglass PICK is a very limited form of PICK compared to PI/OPEN flavour pick and Universe.

Some PICK flavours have evolved and developed useful tools to aid application development and some haven't.

Scott

Gene

I have looked at the Pick data model databases - Pick Universe and Jbase

They all have a similar object library. This means it is labourious

using visual basic to design forms and then connecting to the database.

This was like OI 1.0.

There does not appear to be any forms designer as good as OI at this

stage but I am still looking.

Jack Murray

Ahhhhh yes… Universe. Been on Universe (on a UNIX) for over 5 years. Never had one problem. Not a one! All the data files are perfectly intact. All 800+ megs of data files.

Currently testing Universe on an NT system. It seems to have no problems.

Also been on Revelation for over 6 years. Only had 2 problems. And both of them were caused by the "server" crashing (not the Revelation software). that application is about 400+ megs of data files.

Both types of applications are extremely similar. I really can't complain about any one of the two. These applications are disk intensive.

I strongly believe that if you design and index your files wisely, you will have a very efficient system. A heavily indexed file will oviously use up lots of resources when updating that file. And, an under indexed file will cause selecting and reporting to take excessive time and resources too.

Both Revelation and universe linear hashed files are extremely reliable and efficient.

And always choose variable length, never go for fixed length.

View this thread on the forum...