Przemyslaw Czerpak for example have made a lot of works in harbour
will this work be usefull also for xharbour?
So my proposal is :join the project!
While I understand and respect such proposal, it's not as easy as it sounds.
There are some differences in the way the projects are managed as well as
approach to technical issues, that make it undesirable for some of the
developers. Since both projects are run by volunteers, no one can force
these developers to participate if they do not choose to.
Very Thanks for your reply
I hope also in posting of same harbour developer
can i know wich are some differences in the way the projects are
I don't want force nobody bud explain differences will be usefull for
Good starting point would be to read: "The Cathedral and the bazaar"
The Harbour project was originated by Antonio Linares "bazaar" style, but in
few years (IMO) it somehow converted to a rigid environment controlled by
very few individuals, which rejected/ridiculed, etc. practically anything
originating from others (like me f.e.).
As a result I left the project and started the xHarbour project, which I
believe to this day maintains the "Bazaar" style, where practically all
contributions are welcomed. This is based on the belief that even an
imperfect contribution is preferable to no activity, and can generally be
fixed/perfected quickly thereafter.
I believe the difference in management concepts resulted in the Harbour
project becoming pretty much dormant, while xHarbour strived, and grew to
become a progressive and reliable platform for various applications, on
practically any popular O/S.
After many years of practically no creative activity, Przemek decided to
take over the Harbour project and have brought over many of xHarbour
features as well as many other important improvements.
He's indeed an amazing developer but I object to his communication style,
and his tendency to take control in a manner that discourages other
While it *appears* that the Harbour project is now very active again, it
actually is all based on the work of Przemek, which in my opinion is very
risky for an open source project, as it happened before with Ryszard, which
in my opinion is just as a great developer as Przemek.
Finally there also some differences in technical approach. I believe that
Przemek prefers scientific correctness, while I prefer practical usage.
What may be a good example of such difference is his implementation of AS
<Type> in Harbour OOP. Przemek seems to think that NIL is *not* compatible
with pre-declared types, while I believe that NIL must be acceptable value
for all declared types. IOW Przemek implementation raises error if NIL is
assigned to a property declared as CODEBLOCK, while I believe this must be
allowed. I also believe that Clipper foundation clearly works that way too.
Another recent example is Przemek objection to Extended Codeblcoks saying
they can never be properly supported in PP. After I found away to add such
support to PP he decided to add it to Harbour too. I'm convinced that if it
was in the Harbour project alone, no-one would have had the willingness to
upload such hack against his strong original objections.
> I don't want force nobody bud explain differences will be useful for
> a bridge
What is the future of RMDBFCDX if Przemek has left xHarbour and devoted
himself to Harbour instead?
This is a nice question. The funny think is that RMDBFCDX doesnot included
in Harbour (even if Przemek is one of the main developers of Harbour) and
its included in xHarbour.com. I think that RMDBFCDX need some optimizations
(to be more clipper compatible with clipper in speed) Ordkeyno,
Ordkeycount, filter with logical expressions, caching etc and i dont now
who will make this enhancements! Also there is a need for full kernel
support for multithreaded database applications (somethink that its
allready in progress in Harbour).
> What is the future of RMDBFCDX if Przemek has left xHarbour and devoted
> himself to Harbour instead?
As Przemek is spending more time on Harbour now, does not mean he
"left" xHarbour. I'm confident that Przemek will be fixing any bug we
encounter in RMDBFCDX. At this time we don't have an open bug report
> This is a nice question. The funny think is that RMDBFCDX doesnot included
> in Harbour (even if Przemek is one of the main developers of Harbour)
That's because RMDBFCDX's source is copyrighted xHarbour.com Inc. code.
Does the Przemek have not "left" xHarbour ?
I found a article from Harbour-Dev and like this:
"I do not want to create a flame war here so I will not answer for
any other things then planed or not Harbour code modifications.
Now I answered because I believe that some of my arguments maybe
important in decision about farther Harbour modifications for
other Harbour developers."
Both xHarbour.org and Harbour.org projects are driven by volunteers. So
we must respect their decisions.
I do not believe that Przemek will not help us if we call in his help
to fix specific code. So far, Miguel Angel Marchuet is doing a great
job in syncing RDD code.
> I do not believe that Przemek will not help us if we call in his help to fix
> specific code.
It already happened regarding the infamous "Message not found" bug, do you
EMAG Software Homepage: http://www.emagsoftware.it
The EMG's ZX-Spectrum Page: http://www.emagsoftware.it/spectrum
The Best of Spectrum Games: http://www.emagsoftware.it/tbosg
The EMG Music page: http://www.emagsoftware.it/emgmusic
> Hello Randal,
>> What is the future of RMDBFCDX if Przemek has left xHarbour and devoted
>> himself to Harbour instead?
> As Przemek is spending more time on Harbour now, does not mean he
> "left" xHarbour. I'm confident that Przemek will be fixing any bug we
> encounter in RMDBFCDX. At this time we don't have an open bug report
> for RMDBFCDX.
Maybee they are not bugs but its needed optimizations. As you remember
months ago Przemek confirm some thinks (in some situations CLIPPER is much
faster in filters than xharbour RMDBFCDX) and he said that he dont have
time to do it. If you find this messages you'll remember the situation
(caching, Ordkeyno, ordkeyount, logical expressions must have == .t., etc
> Maybee they are not bugs but its needed optimizations. As you remember
> months ago Przemek confirm some thinks (in some situations CLIPPER is much
> faster in filters than xharbour RMDBFCDX) and he said that he dont have
> time to do it. If you find this messages you'll remember the situation
> (caching, Ordkeyno, ordkeyount, logical expressions must have == .t., etc
Sorry, but this is a distortion of his reply. AFAIR he replied that RMDBFCDX
is fater than Clipper, and the differences you reported were due to caching
which only affecs a SPECIAL CASE of a test.
As i said "IN SOME SITUATIONS clipper is much faster in filters"
Even if someone read the following answer from Przemek he will understand
what i'm saying.
>xHarbour does not use any caches. As long as you are using small
>tables and index so Clipper can keep all index pages in cache it
>will have to be faster on system with fatal IO performance like
>Windows. Additionally when you copy data to file server and the
>network is slow then the difference will be much bigger. In next
>year I hope that I'll find some time to implement global caching
>system and this will dramatically change the performance. To this
>time I can only suggest migration to serious operating systems ;-)
>F.e. in Linux your test compiled by xHarbour is four times faster
>then compiled by Clipper.
> Somewhere in the future it will be done in the way as it should from beginning.
>Patrick said me that you made some changes and xHarbour have now almost the same
>speed as clipper (Ordkeyxxxx - dbfilter) and ask me if its possible to
>check it again. I fully agree about clipper and caching system but this
>have nothing to do with my customers. 95% percent for normal users using
>windows so they dont care about other serious OS. As for the size of the
>sample file is 30mb with 55.000 thousand records (even with a bigger
>sample file i have almost the same results 55Mb). If you think that this
>number is small for normal non client server applications then you are
>wrong. As for the PC it Core Duo 6600 with a gigabit NIC and communicate
>with managed swithc level 2 with a dual xeon server with raid 5 controller
>a 5 SCSI disks. As for what said about linux i allready made some tests
>using Terminal server on a Windows fileserver with excellent results (but
>even that way xHarbour is slower on that operations). There is allready an
>excellent lib (at least for clipper) from OTC that make almost what you
>said but for windows. The results from my test with clipper when i test it
>was excellent. http://www.otc.pl/index.asp?s=22&l=2
What can we do for having a great developer as Przemek in xharbour
Why double the effort for having either half product?
IMO is not good having this differences for same little tecnical
The user that have made more upload/works in each area will decide
most important decision
Sorry Massimo, but you are dreaming.... 80% of software developers are
narcisist, with big self-esteem.... on the floor I work in, there are
same you can't even question their code !
We must just have the idea that there are two languages available,
xHarbour and Harbour. They share the same ancestor, have big
similarities, both have pro and cons and peculiarities.
You write code INTO one of these languages. You bind to it....
I believe that in the short term development will continue divided....
Just as an example, afaik Przemek is linux oriented, Ron is obviously
PS: this message may sound like a flame.... it isn't !
> I believe that in the short term development will continue divided....
IMO is a disadvantage for x/harbour comunity
> Just as an example, afaik Przemek is linux oriented, Ron is obviously
> windows oriented.....
This is not a problem, for all multiplatform product?
or you think that the problem is due to relationship between the
innovative Przemek and conservative Ron?
> or you think that the problem is due to relationship between the
> innovative Przemek and conservative Ron?
Sorry to jump in, but the reason WHY Ron started the xHarbour project
was because Harbour was going a more conservative route. Ron wanted to
add more and more modern extensions to Harbour though was axed by the
Harbour developers just because of the conservative route they wanted
to follow. It's only just now since Przemek took over that they started
to add those extensions. Since than both projects got to look more
This is no criticism, just observation :)