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 : MicroStrategy Inc. (MSTR) -- Ignore unavailable to you. Want to Upgrade?

To: sadsack who wrote (326)	10/7/1999 10:25:00 AM
From: treetopflier	Read Replies (1) \| Respond to of 717

Since u r up 25% since the start of our dialogue who am I to argue... It isn't really an argument at all anyway. I've seen MSTR generated queries against quite large datasets, e.g. Daytons. Numerous problems in query complexity, appropriateness of indexes, query construction, methods of access (query parallelism vs. parallelism through multiple queries). Often the OLTP database isn't correctly constructed to accomodate queries of this nature, or should I say timely completion of these queries. If you have 24 hours to run the query against the working set of 1B transactions, no problem. But Saylor implies real time... Daytons may be old news, but the fundamental problems in working with large datasets are no different then than now. Oracle has added partitioning, improved the optimizer and the performance of parallel query as an example, but to get a query of the nature you describe to run correctly requires that the query leave the realm of 'ad hoc' or generated and become a carefully constructed and 'tuned' query. As soon as it leaves this realm, that of the 'ad hoc' ROLAP, it can and should become part of the application. Once it takes on that character, I don't need Saylor's tools at all. So here I sit with $200K in MSTR licenses and all my BI queries have been added to the core app using the native customization tools that they came with or were built in-house with. Now they are deployed via Java with embedded SQL and PL/SQL and all processes are taking place on an immensely more scalable UNIX box (vs. an NT front end/UNIX back end) with process to process communications via memory between the database and application program. I clearly understand how he is positioned and what the 'frame' does. It obviates the need for the customer to define the presentation interface. Saylor wants to be to 'data analysis' what Gates was to the computer -- in essence a drop in portal or window. No argument. But just like Windows, I wouldn't try to use it for anything large in terms of data size or number of transactions. If you want to build small to medium size applications with small to medium size working datasets you can probably meet with success using these products. Beyond that I'd go elsewhere for solutions. When query duration starts to run into the minutes and hours I'll take a product like Business Objects over MSTR as an example. It really boils down to me feeling like Saylor is standing at the water's edge proclaiming he can part the waters for all to pass and his developers and consultants are left hoping the waters recede a bit so the customers will only get wet to the neck... $70 today. Waters must be receding. ttf