Report Generator 2

[Roger Seiler]

Following the prior discussion about open source development of a report generator, or a fully featured browse class, etc...

This discussion about open source add-on tools for FWH is interesting in view of some springboard projects ripe for open-source development that have been presented here and ignored in the past.

An excellent example is Tsbrowse (for which James Bott cleaned up ver. 6.1 and makes available for download on the Programmers Page of his website). Tsbrowse is a great "springboard" for continuing open source development. It combines Wbrowse, Tcbrowse (much enhanced over the standard!!!) and Tccolumn into a single class with some added features. It works. It is viable. Yet it has been ignored as a platform for robust continuing open source development.

Why?

There are many other examples of open source "springboard" tools that have been offered to the FWH community, but ignored. By "springboard" I mean an FWH class that has been offered that does an important job, that works, but which needs a lot more development to make it fully robust and to keep it up-to-date.

Why are they ignored by the FWH community?

"The World Is Flat" by Thomas Friedman is a best selling book that came out in 2005. It describes how globilization works, including open source development. Compared to the description of robust open source development in Friedman's book, some FWH open source projects seem broken.

Why can't we make open source development work well in add-on FWH tools?

Yet the continued development of Harbour and xHarbour seem to be viable open source projects operating in the way Friedman describes. (But why are there 2 instead of just 1??? Is the original difference in philosophy between Harbour and xHarbour still relevant?)

The "Why" questions probably need answering and solutions found before many will be willing to devote serious time to the kind of open source projects discussed in the prior thread. And there are viable "springboards" already available for an open source report generator, an ad-hoc browse builder, an adhoc query generator, etc., if it would be worthwhile for their authors to contribute them.

However, I can fully understand why Antonio may not wish to get involved in these open source add-on projects because he has his hands full concentrating on building and maintaining the multiple underlying platforms that he provides. So I guess this is mainly a discussion for the rest of us.

Just my two cents. - Roger

[Otto]

>But why are there 2 instead of just 1??? Is the original difference in philosophy between Harbour and xHarbour still relevant?
... and worth that such a small community as “Clipper32” can waste so much resources for developing and maintaining 2 languages.
Regards
Otto

[Roger Seiler]

On Friday, Jan 26th, Tim Stone entered the following comment buried far down in the "Report Generator" subject. To make his insightful comments more visible, I'm taking the liberty of replicating them here...

From Tim Stone:

Reinaldo,

I worked closely with Rene testing the GX Reports and then his follow up project which still exists. I think its safe to say he was burned out with some unpleasant experiences "marketing" GX and so he moved on to the Visual Studio crowd hoping for a better audience for his products. I found GX to be stable and encouraged him to bring the same features to 32 bit.

I think its fair to say he was frustrated with the same type of line I often see on this forum which is "we need something for free." I'm not against Open Source projects, but I am amazed at how often people assume they should get professional products without paying. Most of us know we can't live in a home, eat a good diet, and care for our families without an income. Sometimes the idea that "it should be free" is quickly translated to "its ok to pass it around and not pay the programmer." When that happens, the 3rd party vendors finally walk away.

Over the years ( and I've seen many ... I've been writing programs since CP/M ) I've seen a lot of excellent 3rd party products be terminated because the developers could not earn enough to continue with their work.

I like the idea of a report generator that my clients can use. However, I do all of the reports in my software in my editor and they work fine. After all, reports are really data, not publications, so they are not hard to design using the tPrinter class.

- Tim Stone

[Roger Seiler]

Perhaps another problem with making 3rd party development of add-on tools for FWH, for purchase, is that the size of the potential market hasn't yet reached the critical mass necessary to provide enough orders to make it worthwhile.

In this context, perhaps Open Source development of a few important tools (report generator, browse, query, etc.) would make sense if there was some organizational framework for it. In the past, there has been informal open source development that has provided a few tools up to a point. (for example: Tcbrowse, Tsbrowse, etc.) But continuity has been lacking.

For example, Tcbrowse & Tccolumn and their supporting C routines were origninally contributed by Selim Antar and Harry Van Tassell in 1995. This browse ensemble is currently provided with FWH, but this standard version is very little advanced from the original. On the sidelines, some developers have greatly improved the Tcbrowse ensemble so that whereas the standard FWH version is about 48,000 bytes, the much improved version is over 181,000 bytes with countless additions of functionality. As far as I know, only 6 people are using this much enhanced version, and most other Fivewinners aren't even aware of it. A few years ago, Manuel Mercado combined Wbrowse, Tcbrowse (as it existed then) and Tccolumn into a single class, made it available and then went on to other things. (James Bott cleaned up ver. 6.1 somewhat.) However, the current version of Tsbrowse is missing many of the more recent improvements in the 181K version of Tcbrowse. Clearly, the coherent Open Source development methods used so successfully with Apache, OpenOffice.org, Gimp and other products are not being used to continue development of our add-on tools. Is this to our detriment?

Certainly Antonio cannot be expected to provide this needed continuity for add-on tools development because his efforts have to be focused on the huge effort involved in continuing to advance the development of the underlying FWH engine, which he is doing brilliantly.

So what do you think? Should we try to apply coherent Open Source development methods to establish continuing development of FWH tools like a browse ensemble, report generator, adhoc query tool?

- Roger