Very informative post found on Yahoo (re: Interbase):
Interbase in context
by: yes_know (M/Silver Spring, MD) 12/22/1999 11:56 pm EST
Msg: 25017 of 25055
Interbase is an excellent product, even leading edge in some ways; its problem has always been that Borland is not primarily a database company. It is a company which makes the tools to write applications, and software to manage, deploy and provide security for applications. Because of the nature of their core business, and as a survival strategy during difficult times, they needed to cooperate with a large number of companies. In this context, interbase was marketed as a niche product - a database to be embedded with apps - and not a competitor for Oracle, MS, Sybase etc.
With the development of IB 6, which has replication, and the recent linux popularity, I suspect the IB management team was pushing to make IB a more central product. The problem is that, as advanced as it is, it is based on what is now an older technology - that of strictly relational databases. In addition, it clearly takes a fair amount of effort to develop IB for each of the platforms it runs on, as indicated by the fact that it takes quite a while to syncronize the version numbers.
This is a problem for IB because Borland, which never ceases to amaze me, has recently developed a second database product, JDatastore. I didn't pay much attention to it when it came out, thinking it was a simple lightweight db to bundle with Jbuilder. In fact, it is a very powerful, heavy duty Object/Relational database written in Java. You can read about it at:
sys-con.com
I haven't used it yet, so weigh this accordingly, but it is a much better solution to the problems Borland tries to solve. First, it is written in Java and will run under any up to date compliant JVM. This means, sooner or later, Windows, Linux, OS/2, Mac OS/9, BE OS, whatever. The economics of this should be apparent. Second, it is international, capable of displaying multiple languages simultaneously, including Japanese. Not only can it sit between an even larger Oracle db and the application, acting as a kind of cache, it can also handle the conflict resolution resulting from conflicting updates to the data, apparently incorporating Borland's undervalued (why do those words always go together) Midas technology. It stores data as streams, allowing it to store objects of any sort, not just tables, and it contains its own file system for accessing these objects.
JDatastore is clearly the database for the future, not IB, and this was probably the issue which caused the development team to leave. IB is still a profitable, excellent product -which I use - but destined never to be the star it could have been.
A final comment: Borland is a company with an embarassment of technological riches, a dedicated development staff, a lean, consolidated development method, millions of committed developers worldwide and an international base of customers, no debt and enough money to execute. With the recent surge in the stock price (3-5 times what it was until 6 weeks ago) they have the currency of options available to motivate employees further. I suspect that the major stockholders are looking at this as a huge opportunity, and I don't think any company can offer them a buyout to match. Those of you who have some technological knowledge and are not focused on the next few days or weeks probably think this is increasingly obvious, but I'm interested in informed responses. I have no special knowledge or contacts, so this is just derived from the usual sources. |
