Innovation in IT

TalkBack 1 of 17:: Next »

Thread View
Flat View

A great deal of activity but how much achieved?: It may well be that there is a great deal of activity going on but how much of it is really innovative?

Hardware is getting faster, that is agreed but I question whether this is really influencing software very much.

The main product of open source so far seems to have been cheaper versions of existing technologies. This isn't in itself a bad thing but it is hardly innovative.

Greater activity in programming won't necessarily lead to greater innovation. Software is after all a branch of applied mathematics. If Turing and Von Neumann had merely been engineers rather than having a profound insight into the theory of algorithms and logic then the computer as we know it today probably wouldn't exist. The software world seems to have lost sight of its mathematical legacy and this mathematical legacy is the chief wellspring of innovation in software. To give a new take on the monkeys writing Shakespeare, if you put a thousand programmers in front of a thousand computers, how long would it be before they stumbled upon the relational model of data by accident? Whether North America, Europe of the far east wins in the software world depends primarily on the quality of their education systems, in particulary in the field of mathematics.

As regards integrating databases and applications. Fifteen years ago I was using the Wang PACE product. Aside from having full declarative referential integrity (something Oracle got in version 8 and Sybase users are still waiting for) the development tool generated the whole application from the metadata. This wasn't just individual screens but the whole navigation between screens based on the foreign key references defined in the database. PACE also had host language triggers, something that is now regarded as an "innovation" by Oracle and Microsoft, but is, when you stop and think about it actually regressive. Why replace a declarative language like SQL with a procedural language like Java or C#. This is (in every sense of the word) loopy.

Your comment on SQL seems to be repeating a common fallacy and a very widespread misunderstanding about the relational model. SQL doesn't need to be replaced because volumes of data will get bigger. SQL needs to be replaced because it is logically inconsistent. The whole point of the relational model is that the logical and physical are separate. Provided the logical layer remains the same, how you represent the data physically is totally irrelevant. Changing language therefore for a properly implemented RDBMS should make absolutely no difference to performance. If changing language does improve performance the RDBMS is a bad implementation of the model and you should be looking for a replacement!; Posted by: jorwell Posted on: 07/11/05 You are currently: a Guest | Members login | Terms of Use