A major advantage of 32-bit OI seems to me to be in marketing terms. When prospective clients' IT folk are present "Is it 32-bit?" has been a frequent question and the answer "No" has been greeted by derisory looks from them. The fact that 16-bit has proved robust and to perform acceptably seems in such circles to have been a secondary issue. No longer!

However, most of our existing clients have to find upgrade fees, etc. from departmental (rather than corporate IT) budgets. If the upgrade is free then they'll accept it gratefully. Otherwise, if their 16-bit OI system has served them well, they need to be convinced of sufficient real-life advantages to justify the cost.

I've looked through the announcements, etc. on 32-bit OI and it seems (rightly, no doubt) to be hailed as the best thing since the proverbial … However, I have found nowhere any sumnmary or description as to why it is the best thing … and what advantages existing users would enjoy from the upgrade. Sprezz tell us that record length will go way above the present 64k (whereas the longest any of our users has generated has been less than 10k (including lots and lots of text notes). I can see longer records could be useful in indexes, quickdexes, etc. - but this limitation hasn't yet been a problem.

Can anyone summarise the case I can make to my clients' end user departments (and IT depts, for that matter) for upgrading to 32-bit, especially if this involves money?

Thanks

Richard Guise

Tornado Property Systems Ltd.

Richard,

If you send me an email at the link provided below I can send you the two 32-bit OpenInsight email reports that we have already produced for the sake of keeping our clients and business partners informed of the progress of this product. In addition to summarizing some of the technical improvements (e.g. removal of the 64K barrier) it also explains what the costs will be for existing OpenInsight developers and end users.

You bring up the old question of "If it ain't broke, why fix it?" We still have many clients running AREV for just this reason. In many cases, due to the nature of the application, there really is no compelling reason to move them to OI (16bit or 32bit). Now, if the clients decide to start upgrading their OS and DOS support is eventually eliminated (as always threatened by Microsoft) then we now have a reason to migrate. I believe the decision to move off of 16-bit OI will be made along similar lines.

We are encouraging as many clients to upgrade because the performance on NT/2000/XP systems will be much improved. We also support OI applications running on Win2000 Terminal Server. However, as you may know, NT-based operating systems must load additional programs (NTVDM and WOW) in order to run 16-bit applications. This puts a great burden on the system, especially when multiple users start accessing the application. OI32 eliminates this problem.

So, your customers may not need to upgrade if they are having no problems with 16-bit OpenInsight. But get as much information as you can to help them make the best decision.

dbakke@srpcs.com

SRP Computer Solutions, Inc.

Don,

…it also explains what the costs will be for existing OpenInsight developers and end users.

Interesting. I've never seen any pricing info from Revelation. Please mail me too.

- Oystein -

Øystein Reigem, Humanities Information Technologies, Allegt 27, N-5007 Bergen, Norway. Tel: +47 55 58 32 42. Fax: +47 55 58 94 70. E-mail: oystein.reigem@hit.uib.no. Home tel/fax: +47 56 14 06 11.

Richard-

I agree with your assessment that if the 16-bit version of OI worked well for clients, and is still working well, there is probably no compelling reason for a user to switch to OI32. The 32-bit world didn't just appear- it's been with us since Windows 95 and OI has run well on every Windows platform since.

So why switch?

First, there is that marketing angle. There are a number of developers who can't get their software installed at a site because the IT department won't allow 16-bit applications on 'their' network. The 32-bit version gets OpenInsight past that hurdle automatically.

Secondly, in the 32-bit world it is easier to interact with other 32-bit applications. For example, if an OI application is using our Client/Server toolset, the developer had to put a section in the ODBC.INI to thunk the drivers down to 16-bit. In some cases the developer would have to use 16-bit drivers, and they are becoming fewer and further between. With our 32-bit OpenInsight a developer doesn't need any special changes.

Thirdly, a 32-bit version of OpenInsight allows us, Revelation, to provide our developers and users with more options. We're working on integrating other technologies, such as ActiveX/OCX, into OpenInsight. We couldn't do that in the 16-bit world.

Fourthly, (is that a word?)a 32-bit version has allowed us to break some of the internal limits in OpenInsight. You may not have any records larger than 10k in the application that you mention, but have you ever written or run a report that loops through data and assembles strings on the fly? In such a situation it is fairly easy to hit the 64k limit. Now that limit is gone. People would shy away from relational indexes in the past because ther was the chance that you could try to shove more than 64k worth of data into a record. No longer a problem. You couldn't have dictionaries with many fields because the %fields% control record would try to exceed 64k. No problem. How about a btree.extract call that had more than 64k worth of results? Not stopping us now. And everything runs faster!

Finally, (there are plenty of other reasons, I'm just tired of typing) with the 32-bit version of OpenInsight we can open up to settings and environments that we couldn't have in the 16-bit world. Unicode, for example- OI will be able to run on Windows versions that use non-roman characters. We can store images, sound bytes, and video in linear hash files. We can store native XML data- just use symbolic fields to extract the data. Linux is also a possibility.

As to pricing, it will be this:

Works members get the 32-bit Development upgrade for free.

New OI 32-bit will be priced about the same as OI 16-bit

Upgrades to existing systems can be had for 90-days after release (scheduled for 15 February) for 55% off of the price of the user count. For example, if you have a 5 user system, and five users usually cost you $1000, you can upgrade them to OI 32-bit for $450.

We'll be sending more information out via mail and email( if you don't receives email form us and would like to please send an email to Info@revelation.com), as well as here on the website.

Thanks-

Mike Ruane

Revelation Software

Mike & Donald

Many thanks indeed. I'm sure I won't be by any means alone in finding your responses very useful indeed.

I can see us deploying all new installations on 32-bit and redeploying (or upgrading) all single-user installations. Smaller network users (e.g. up to, say, 5 users) should find the cost acceptable.

However, with a cost of, say, $2,000 or more, my sort of client needs to see benefits which he can understand, which are relevant and which are substantial enough to justify the price tag from his point of view. I think the arguments would impress the corporate IT folk (who also have quite generous budgets) - but my systems tend to be "departmental" rather than "corporate" and therefore in the hands of end users who are non-technical and have rather tight budgets.

We'll see! However, from my viewpoint I'm looking forward to the upgrade.

Thanks again

Richard

Richard,

The great thing about this upgrade is that you will have minimal or NO work to do on your application. In many cases developers may be able to support both OI32 and OIv3.75 off the same code base.

As others have said - this new OI32 is what we have all been waiting for for years! My impressions are that my OI applications run about twice as fast - just by dropping the application into OI32 - no changes at all.

Richard Bright

r.bright@ark.co.nz

Richard,

My impressions are that my OI applications run about twice as fast

What configuration are you testing this on? Are your applications primarily UI, I/O, or a mixture of both? My reasons for asking is that we are getting some improvement but not "twice as fast".

dbakke@srpcs.com

SRP Computer Solutions, Inc.

Was I exagerating? Maybe.. Maybe only a little. I did some testing as a result of your challenge and am happy that my observation wasnt wildly off the mark. Sufficient to say that some things were dramaticly faster, other things perceptably faster but not something I could quantify.

Platform win2000 Pro P2 350mhz / 512 Ram / SCSI HD - running local.

and on other PC (winXP Pro, 1.5Ghz / 512 DDR / 10K HD) OI32 is a screamer! So cool ….

Richard

and I'll vouch for a similar improvement on

P3 650 with NT 4.

Simon

Rebus HR

Don't tell me we are finally getting 'LightSpeed'?

Don't say that!!!

David,

Funny about your comment … I thought about it at the time I made my original post and decided it had bad associations.

Interestingly, in the evolution of OI32, little was made of the speed improvement - the focus has been on breaking the 64k barrier - fixing long standing bugs in Btree.Extract, compatability with other 32bit dlls, improved stability etc - getting the product right. Any mention about speed was a bland comment about 'might appear slightly quicker'.

But for those who are using a full 32bit system (Win2000 Pro or WinXP) the speed improvement quite distictive. Heck, most would have been delighted to have the 32bit version even if there was a small performance hit. But here we have a speed bonus …

Richard Bright

View this thread on the forum...