What makes a 5GL?
Impossible, but ...
Comment From time to time vendors in the application development space have claimed 5GL (5th generation language) capabilities. If you think about this for a moment you will realise that this can't be true. The idea of GLs is that each is an abstraction of the former, so we had machine code (on which I cut my teeth as a developer); assembler; third generation languages such as COBOL, Java, C++ and so forth; and finally 4GLs (now sometimes referred to as ABLs-advanced business languages-as 4GL seems to be out of fashion).
You can't get a further level of abstraction than a 4GL so there can't be such a thing as a 5GL, just as you can't abstract further than meta-meta-metadata (which is what high-end repositories provide).
So, if you can't get beyond a 4GL, what can you do to a 4GL to make it a quantum step forward compared to current 4GLs, even if it isn't actually a 5GL?
One possible answer is that you could build a data federation engine into the development environment. This would allow you to create applications that accessed diverse, heterogeneous data sources. Now, data federation has typically been thought of as being used for EII (enterprise information integration) and the premise has been that it is about queries. But a database look-up is, from a theoretical perspective, just a query, even if it is going to be used for a transactional application, so supporting data federation within a development environment sense. Indeed, it is easy to see how such an approach could be used for building MDM (master data management) applications, for example.
So, is there anybody actually doing this? Well, as you might guess, yes. A small UK company called Abbro Interactive has a tool called Abbro (now in version 3-it has been in use for some seven years), which does exactly this. You can go to the company's website and take a look at the facilities provided but the key to note is that this is a 4GL with extras that, in the case of federation, includes the ability to reverse engineer existing databases to create database views that can then be merged or joined, and caching capabilities so that reading the same data from these sources does not mean repeated database access.
The product is based on its own scripting language which is interpreted dynamically at run-time by the Abbro engine. It has to work this way because it is intended that applications will be event-driven, so event exits may occur at any time. Note that this facilitates the deployment of workflow as well as alerts, notifications and so on. It includes the ability to scan documents and populate forms therefrom, support for bar codes (and, at least in theory, RFID tags) and GPS messages so that you can determine the locations of things.
As I said: a 4GL with extras.
However, you can't buy Abbro as a product: at present all you can do is to have the company build an application for you using Abbro (based on a proof of concept, if required), which should be significantly faster (in terms of delivery) and cost less money than would normally be the case. Once you've got the application you can customise it but the underlying logic will be developed by Abbro Interactive for you. The company does have one package built on Abbro, for export documentation, but its potential uses are much broader than this. The company is considering how it might make use of channel partnerships but however good the product is, we cannot expect to see widespread deployment while Abbro itself remains the only company doing core development work.
Copyright © 2007, IT-Analysis.com
Sponsored: Are DLP and DTP still an issue?