Re: ISQL forms on NT?

walt@wst.b30.ingr.com (Walter S. Tyszka) wrote: :Nils, : :Perhaps you are missing another point here? A lot of graphics intensive :businesses are retiring their old UNIX systems in favor of newer, more :powerful, usually less expensive, Intel Pentium (or Pentium Pro) multi- :processor systems running Windows NT. They could run Unix as well (SCO Openserver, Unixware or Solaris X86) at more or less the same price. I do however understand that they actually do not, and many don't want to for more or less good reasons. : The primary focus of these companies :is not databases, or database applications, (which historically have been :well implemented on single processor multiuser system) but graphics (which :historically excel on a multiprocessor single user environment). The :fact that most of these graphics intensive applications make use of RDBMS :systems was not much of a problem in a UNIX environment, but is becoming a :real stumbling block in the Windows NT world. The end user applications are graphical. The databases (of the type we talk about) will normally be run on a separate server. Whether that runs Unix or Windows NT doesn't make much of a difference to the applications. Whether you use connectivity software from Informix or an ODBC driver what you install on the client is the same software in both cases. The management issues with a Unix server is however a problem. There are simply too few Unix gurus in the world to cover the rapidly expanding market for this type of servers. Unix's expert friendliness is of no help if there is no expert available. :The end users are simply replacing older semi-proprietary UNIX systems with :newer "open" systems which they can purchase from a multitude of vendors. To call Windows NT more "open" than Unix is somewhat strange. The multitude of vendors must be hardware vendors and any hardware that can be used with Windows NT can generally also run Unix. :Many end users ran an OSF/Motif X Windows "environment" on their UNIX :operating system - primarily in a "text within a window" or tty mode rather :than a pure GUI mode. Now they are migrating to a Microsoft Windows :"environment" on a Windows NT operating system (which looks very similar :to the OSF/Motif X Windows "environment" which they where used to). Very few of these users used a GUI on Unix. Most where character based terminal applications. (It doesn't make much of a difference to our arguments though.) : The :graphics intensive applications which they run usually create their own :window to run in, or are full-screen windowless applications, and the user :might not care (or even know) which underlying OS is in use. :In a homogeneous environment of non-UNIX platforms it is wonderful to have :Informix ESQL/C, ODS, XPS, and IUS available; but it is a tragedy to loose :I-SQL, and I-4GL capabilities. Yes. The problem with I-SQL is that the tool is so old. Although I don't know for sure I doubt a little that the number of users of it, and the number of those who are interested in moving to Windows NT *and* continue to run the applications there, is large enough to warant a port (by Informix) or a new tool (by someone else). If I am wrong why haven't someone done it? As to 4GL Informix tried with 4GL/GX (I believe that was what they called it) and it failed misserably. One major reason was the bad implementation. However most users also understood that one realy have to rewrite an application to run in an event driven GUI. Informix's answer is NewEra. The fact that it's hard to convert from 4GL to NewEra is not Informix's fault, but the inherent impossibility of creating a truly event driven interface using the 4GL input statement and some other 4GL language constructs. Noone have done it, nor will ever be able to. There is a compiler available that does make your 4GL programs into a semi event driven application, and as far as I know it is even available to do exactly what you want. Namly run your application entirely on Windows NT with MS Windows based PC's as workstations. This is usefull for the kind of stopgap solutions you are talking about. It is however in no way a viable solution for long term GUI development. :The DBMS vendor which makes this migration from UNIX to Windows NT the least :painful to the end user will probably win the lion's share of the DBMS :software sales in this arena. You miss the mark completely here (unless you by "this arena" means only those who now run I-SQL/4GL applications on Unix - but that's a small market compared to what the Windows NT database market is going to be). 1. Only prior Informix users will be interested in I-SQL/4GL on Windows, and then again only a small percentage of them. The rest will create or buy new applications. 2. Most DBMS users on Windows NT will use applications written for the purpose - true GUI applications or web server based applications. That market is wastly much larger than the one created by people converting from Unix and wanting to bring their character based applications with them. :Informix has done a tremendous job in porting :world class dataservers to this environment, but have done very little to :support their long time I-SQL, and I-4GL users. Despite all I have said this is a problem for Informix. The first time they did a Windows NT port (SE for NT) they hoped to sell with no tool available but ESQL/C. Utterly stupid, and it didn't sell that much either. If 4GL fow MS Windows had been available at the time in a well done implementation and they had done a top job of supporting tools like Visual Basic (I don't think Delphi was out yet) the situation would have been quite different. Then they tried with 4GL for MS Windows. Very badly implemented and not usable. As proven by others it could have been done much better and was realy very interesting at the time. Then they tried again with NewEra but didn't have a clue about the requirements of the MS Windows developers. They came from a Unix world where few developers had heard about high quality GUI's with all kinds of bells and wissels that are so important to make good programs for the end users. The key issue at the time where support of VBX's. Very few people inside Informix even knew what that was, much less how they where used or why they where required. Now finally with NewEra ver. 3.0 may be, just may be, they have a usefull product. It's late though, and to even have a chance they should reduce the price to well below US$1000. More interesting is their work with Java. If that is realy well executed (which is very hard indeed, but they may be on the right track) they may become leaders. :I am certainly not trying to make this a UNIX vs Windows NT issue Neither am I. : (because I _love_ UNIX); I do not "love" any platform. I try to use the best one for whatever my purposes are. If one start loving things too much in our bussiness one can too easily get into all kind of problems. The socalled religious wars we see from time to time are so destructive for everyone involved. : but Windows NT is making a serious entry into e