Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : IFMX - Investment Discussion -- Ignore unavailable to you. Want to Upgrade?

To: investorgal who wrote (10109)	3/31/1998 3:09:00 AM
From: Charles Hughes	Respond to of 14631

I like the tagline too. I know where @jim is coming from though - there have been a slew of products in the past with similar taglines, and nobody cared. Mostly upgradeable products nobody ever upgraded - like the Intel overdrive PCs (they stung me for a couple hundred on those.) Who would care: engineers, DBAs, and others whose livelihood depends on anticipating needs, future required job skills, and development trends. For us it says it all. Cheers, Chaz

To: investorgal who wrote (10109)	3/31/1998 9:56:00 AM
From: Mark Finger	Read Replies (1) \| Respond to of 14631

investorgal, It appears that you have a lot of inside contacts. I still have a couple of questions. First, the 1000 sites was from a December 1997 release, 4-5 months after Phil left/was booted. Therefore, that would have been from Mr. F. In other words, you are implying that F. is continuing the "Phil fudge". I can understand this applying to 1996 sales (very minimal, since they were almost giving away parts, with the "developer's kit" being priced at $125 or so, but not including a full engine), and probably would have been only in the $5-10M range (at best). I can also see the same applying to many of the numbers of implementations in 1H 97, when Phil said several hundred sales. >>As with any new product, IUS did have its stumbles, mostly with the >>datablades. Many of the problems were 3rd parties not writing >>enterprise class code for a large environment (can you say single >>threaded?). This caused R&D to take several of the key blades >>in-house to get them to work properly in the environment IFMX >>targets. The problem was having a scalable engine choke at the >>datablade. I believe now you will see 5 or 6 key datablades >>available thanks to Mr. Saranga and company. There are two parts of this. First, you only need multi-threading on larger or more complex objects. For simple objects, you would not need multi-threading within the object, but just that the code be "re-entrant" and using "thread-safe" libraries (i.e., no global variables, so that two different "threads" can be using the code simultaneously). Even the second case is foreign to the bulk of developers and can take time to learn. The other part of the problem is that performance inside an engine is quite different from the performance in most programs. Just moving data is highly optimized. Further, allocation of data is one of the biggest time consumers in many programs (and the more complex objects), and this will be optimized differently in many engines. This means that there will be a very steep learning curve for Datablades, and that only some of the learning curve will transfer to other products (i.e., to Oracle, IBM and Sybase), because the internal differences will be more than the typical C-programming library. These products will vary more than most developers are used to (even between NT and Unix). In other words, for a year or two, you probably will see a lot of difference in quality of Datablades, with the best ones coming from people who work closely the IFMX developers to really get the details. It will take time for the people who are not that close to get the learning to write the Datablades. Re--the Sabre implementation. If as you say, that implies "to others" that the full integration of IUS back into the DSA will not be as bad as Larry and others have been saying. IFMX has done "one-off" cuts before. Version 8.0 XPS (for MCI) was a similar case, and that was very successful--it gave IFMX a very large reference site in production about the time the real product was being released to other customers. The Sabre case could show the similar scaling capability as IFMX fully integrates the two code lines (DSA/IUS), and they could have the same kind of reference site.