Re: Oracle and Partner Push "Informix to Java" Migration
On Mon, 27 Jul 1998 08:02:47 GMT, afalout@ihug.co.nz (Andrej Falout) wrote: >I was waiting for some of the c.d.i. "Big Boys" to do this, but they >are obviously Sleepless in Seattle. Yes. >If I remember correctly, Nils said in his last post before vanishing >about month ago that he will investigate this ArtInSoft and report >back to group. Nils? All I have done so far is to get a demo of the product and talked a lot to these people. Yes it sure looks like they have a real product, and they know what they are doing. It seems like they are able to generate sensible Java from 4GL, but there may be some limitations depending on your coding practices. At my request they are also doing some changes in the GUI interface that is generated. I expect to be able to test the product within a month or two depending on when we can get it from Artinsoft (also depends on some negotiation of conditions, not so much the product itself). To me it's most interesting for generating Java from all our batch programs. It's a little harder to create usefull Java for programs that contains a lot of 4GL user interface code. Even if they do a sensible generation of proper object for the purpose, I am not so sure it will be as good as creating it all from scratch. Some have said they have a problem in that the Java code is generated as one Java applications so you loose the ability to run on the server and therefore will loose a lot in performance. To the extent they haven't allready solved this I think it can be solved. At any rate it doesn't matter much for our batch applications which can easily be run on the server anyway. We'll see. You should also be aware that all the talk from Oracle about using it to port to an Oracle database isn't particularly interesting. Sure they do have some extra support to make this a little simpler than it would otherwise have been. However if you continue to use the best database engine there is your conversion to Java will be simpler. The impact on Dynamic 4GL may not be all that big. We convert to Java because we intend to sell a large number of our products. In that case we aren't too happy about the runtime price of 4GL, but even more important we can't live with it's limitations in the GUI designs that are inherent in any such tool. We may however decide to use Dynamic 4GL for a time for some purposes. For a shop that has a large inhouse 4GL application or one that they sell to a relatively small number of customers (may be up to several hundred) Dynmaic 4GL may be the best tool to come to market with something that at least looks like a GUI within a reasonable timeframe. The other alternatives seem to be from Querix or Etrus. The main advantage of Querix seems to be their runtime free pricing. Their other listed advantages aren't very impressive to me. As to Etrus it's quite another issue. I am talking to them as well. You may want to have a look at their web site at http://www.etrus.com Etrus does GUI conversion using Java on the client (much better than Tcl/Tk that Dynamic 4GL is currently using - but a Java version is supposed to be in the works) and seems to run quite well. Using Java you can avoide having to install any client side code. You can run your user interface as a Java applet in a browser. For many this is a significant advantage. If you want to install it as an application (to awoide using the browser) that's of course an option as well. Etrus development and runtime prices may also be lower than Informix have on Dynamic 4GL (possibly depending on your particular deal with Informix). All of these new tools (including Dynamic 4GL within a short while) run on Linux by the way. Isn't Linux nice. All of a sudden we have all these alternatives and they all have to think about price. Nils Myklebust NM Data AS Norway E-mail: Nils.Myklebust@nmdata.com FAQ at: http://www.iiug.org/techinfo/faq/faq_top.html (Now with ODBC info under "Third party products".)