This is what I´ve been told to do on my request to fix the lack of memory crashes caused by the latest fractint update. /"You just need to back off on the demands of your fractals// //so they are not at the edge of Fractint's capabilities."/ This means the end of my multifractal project after 12 years, nothing less. So the last Mohawk has to go. Albrecht
Albrecht, I choose my words carefully and always answer you respectfully. The words you are attributing to me are not respectful. Of course I cannot control what you think when you read my emails, that is your responsibility, but best not to confuse those thoughts with my words. I don't believe I ever told you to "back off". The actual situation is that Fractint was built with thousands upon thousands of hours of volunteer time, a large number of them by Jonathan and myself, but also from hours from many others. You have benefited from this contribution, and I hope you have some appreciation. Fractint is built with early 1990's technology that has severe limitations. Endless optimizations have squeezed a huge range of zooming into finite resources and allowed complex parser formulas to be utilized despite old technology. Apparently Fractint supported your multi-fractal project for the last 12 years. I am not exactly sure why suddenly your project must stop because of some unintended consequences of Jonathan's last patch, his first in something like four years. But of course whether you continue your project or not is your decision, and yours only. In the entire time I have worked on Fractint the developers have never put any pressure on each other. We were completely clear that the effort was voluntary and done with discretionary time by people who had many obligations. I'm not going to start now by putting any pressure on Jonathan. As if that would matter. Jonathan is undoubtedly interested in the technical problem posed by his patch and the demands of your fractals. I'm guessing he will look at it. Jonathan has an extraordinarily demanding day job working at a nuclear power plant. His discretionary time is limited. So there is no telling when or if he will get to it, or if he does get to it, if he will make any changes that will relieve the constraints your multifractal project is experiencing. We don't know if the problem we are seeing is caused by Jonathon's patch or if an existing memory violation was changed from being innocuous to fatal. Memory problems are among the most difficult to debug. My suggestion to you is not to "back off", but to see if your researches can continue less close to the edge of Fractint's limitations. Maybe they can't, I don't know. My uninformed guess is that if you put the same effort into optimizing your parser formulas that we put in the fractint coding, you could happily continue. But I might be dead wrong, and I really don't know quite what you are attempting or what your vision is. SO "back off" would seem to be inappropriate advice coming from me. If you really want help from Jonathan, do some experiments and find out exactly what triggers the error. Exploit the fact that he is interested. Give him some information, but respect the fact that his time is limited. But asking me to somehow get Jonathan to do what he either can't or won't is a losing proposition. I have zero authority over Jonathan nor do I want any. I last changed fractint code maybe 15 years ago. Even when I was active my skill set was different than Jonathan's. Right now I am not the answer to your Fractint problems. You have a number of fans in the forum, and have injected some life after Jim Muth's changing interests reduced his participation. Jim has single-handedly kept the forums alive for a very long time. You have made things very lively recently, and a number of folks, including me, appreciate that. That said, all things come to end sometime. When the activity in the Fractint forums ebbs to a trickle, nothing tragic will have happened. The universe will go merrily on. I can hardly believe that there is still a fan base for Fractint after all these years. One of those technological mysteries. Every day interesting material is posted is a bonus. I wish you well. I hope you keep posting, but if you decide not to, your decision will be respected by me. Tim On Thu, Mar 12, 2015 at 6:32 PM, Niekamp <Niekamp@ish.de> wrote:
This is what I´ve been told to do on my request to fix the lack of memory crashes caused by the latest fractint update. /"You just need to back off on the demands of your fractals// //so they are not at the edge of Fractint's capabilities."/ This means the end of my multifractal project after 12 years, nothing less. So the last Mohawk has to go. Albrecht
_______________________________________________ Fractint mailing list Fractint@mailman.xmission.com https://mailman.xmission.com/cgi-bin/mailman/listinfo/fractint
Hear, hear Tim. Cheers and thank you. Nicely put! John W. On 12 March 2015 at 20:08, Timothy Wegner <tim@tswegner.net> wrote:
Albrecht,
I choose my words carefully and always answer you respectfully. The words you are attributing to me are not respectful. Of course I cannot control what you think when you read my emails, that is your responsibility, but best not to confuse those thoughts with my words. I don't believe I ever told you to "back off".
The actual situation is that Fractint was built with thousands upon thousands of hours of volunteer time, a large number of them by Jonathan and myself, but also from hours from many others. You have benefited from this contribution, and I hope you have some appreciation. Fractint is built with early 1990's technology that has severe limitations. Endless optimizations have squeezed a huge range of zooming into finite resources and allowed complex parser formulas to be utilized despite old technology. Apparently Fractint supported your multi-fractal project for the last 12 years. I am not exactly sure why suddenly your project must stop because of some unintended consequences of Jonathan's last patch, his first in something like four years. But of course whether you continue your project or not is your decision, and yours only.
In the entire time I have worked on Fractint the developers have never put any pressure on each other. We were completely clear that the effort was voluntary and done with discretionary time by people who had many obligations. I'm not going to start now by putting any pressure on Jonathan. As if that would matter. Jonathan is undoubtedly interested in the technical problem posed by his patch and the demands of your fractals. I'm guessing he will look at it. Jonathan has an extraordinarily demanding day job working at a nuclear power plant. His discretionary time is limited. So there is no telling when or if he will get to it, or if he does get to it, if he will make any changes that will relieve the constraints your multifractal project is experiencing.
We don't know if the problem we are seeing is caused by Jonathon's patch or if an existing memory violation was changed from being innocuous to fatal. Memory problems are among the most difficult to debug.
My suggestion to you is not to "back off", but to see if your researches can continue less close to the edge of Fractint's limitations. Maybe they can't, I don't know. My uninformed guess is that if you put the same effort into optimizing your parser formulas that we put in the fractint coding, you could happily continue. But I might be dead wrong, and I really don't know quite what you are attempting or what your vision is. SO "back off" would seem to be inappropriate advice coming from me.
If you really want help from Jonathan, do some experiments and find out exactly what triggers the error. Exploit the fact that he is interested. Give him some information, but respect the fact that his time is limited. But asking me to somehow get Jonathan to do what he either can't or won't is a losing proposition. I have zero authority over Jonathan nor do I want any. I last changed fractint code maybe 15 years ago. Even when I was active my skill set was different than Jonathan's. Right now I am not the answer to your Fractint problems.
You have a number of fans in the forum, and have injected some life after Jim Muth's changing interests reduced his participation. Jim has single-handedly kept the forums alive for a very long time. You have made things very lively recently, and a number of folks, including me, appreciate that. That said, all things come to end sometime. When the activity in the Fractint forums ebbs to a trickle, nothing tragic will have happened. The universe will go merrily on.
I can hardly believe that there is still a fan base for Fractint after all these years. One of those technological mysteries. Every day interesting material is posted is a bonus.
I wish you well. I hope you keep posting, but if you decide not to, your decision will be respected by me.
Tim
On Thu, Mar 12, 2015 at 6:32 PM, Niekamp <Niekamp@ish.de> wrote:
This is what I´ve been told to do on my request to fix the lack of memory crashes caused by the latest fractint update. /"You just need to back off on the demands of your fractals// //so they are not at the edge of Fractint's capabilities."/ This means the end of my multifractal project after 12 years, nothing less. So the last Mohawk has to go. Albrecht
_______________________________________________ Fractint mailing list Fractint@mailman.xmission.com https://mailman.xmission.com/cgi-bin/mailman/listinfo/fractint
_______________________________________________ Fractint mailing list Fractint@mailman.xmission.com https://mailman.xmission.com/cgi-bin/mailman/listinfo/fractint
Am 13.03.2015 um 04:08 schrieb Timothy Wegner:
I choose my words carefully and always answer you respectfully. The words you are attributing to me are not respectful. Of course I cannot control what you think when you read my emails, that is your responsibility, but best not to confuse those thoughts with my words. I don't believe I ever told you to "back off". See your message to me 120315 03.20 quoted literally Albrecht
Albrecht Niekamp wrote:
Am 13.03.2015 um 04:08 schrieb Timothy Wegner:
On Thu, Mar 12, 2015 at 6:32 PM, Niekamp <Niekamp@ish.de> wrote:
/"You just need to back off on the demands of your fractals// //so they are not at the edge of Fractint's capabilities."/
I choose my words carefully..... I don't believe I ever told you to "back off".
See your message to me 120315 03.20 quoted literally.
Must have been a private message, for I could not find anything from the List where this was originally said, especially around that date and time. Or did I miss an email from the List?? But if it were private, why make it public?? Sincerely, P.N.L.
Am 13.03.2015 um 04:08 schrieb Timothy Wegner:
Apparently Fractint supported your multi-fractal project for the last 12 years. I am not exactly sure why suddenly your project must stop because of some unintended consequences of Jonathan's last patch, his first in something like four years. The reason is that in the past I could be sure that all improvements were "safe" and allowed going on in the development. That has changed and there is no room for improvements anymore, even if I replace outdated code. This happened when all commands dealing with function conflicts could be deleted with the new version and replaced by the commands to replace default functions - consuming less buffer space. Nevertheless I now must face the lack of memory crashes, apparently already with simple generation - in Jack´s case. ; Believe me i put the same efforts into optimizing the code as Jonathan, otherwise the project had ended long ago. David: "Your work reminds me of just how awesome a piece of programming Fractint is. The people ages back who were astounded at a 640x80x256 image would look at yours and be completely dumbfounded." This "balancing" on the edge is what makes fractint so special, leaving much space for personal creativity and finally is the reason that the project probably is unique worldwide. ; On my "to do" list is just one more item: The "Warp" effect needs better adaption to the participating formulas, sometimes the only effect is the color compression.Just a few lines will be enough. Albrecht
Albrect,
The reason is that in the past I could be sure that all improvements were "safe" and allowed going on in the development.
That is not strictly true. You are using the developmental version and there is always the possibility for some feature being broken. There is clearly no guarantee as stated on the What's new page in Help.
That has changed and there is no room for improvements anymore, even if I replace outdated code. This happened when all commands dealing with function conflicts could be deleted with the new version and replaced by the commands to replace default functions - consuming less buffer space. Nevertheless I now must face the lack of memory crashes, apparently already with simple generation - in Jack´s case.
I did not have any problems with my development environment, which is why I released it. You pointed out a problem, and I have started working on a solution. I will not simply back out the changes, since that will cause Xfractint to not be able to run your formulas, which was the point of the patch in the first place. I need to collect some empirical data comparing formula string size to the final formula memory size and then extrapolate out to the 96K limit of the available extraseg memory for both DOS and Linux since the data structures are of different sizes. There is the caveat that since the extraseg memory is shared memory, it is possible to overrun memory used for other features. You may run into this when switching to a text mode and then back to your image, for instance. I do not, however, need prodding. I work on Fractint as time allows. Having to rebuild my Win XP development environment recently when the hard disk crashed has not helped. Jonathan
Am 13.03.2015 um 23:39 schrieb Jonathan Osuch:
I do not, however, need prodding. I work on Fractint as time allows. Thank you, Jonathan. It´s good to know you´re working on the problem. And I apologize if the impression came into being that I put a pressure on the fractint team. I´m fully aware what a phantastic job you´re doing squeezing such a complex software into 1.441 KB. As there is no feedback sofar I would like to know if the xfractint users are able to generate my images. Thanks again, Albrecht
Albrecht,
As there is no feedback sofar I would like to know if the xfractint users are able to generate my images.
In my efforts, I failed to recognize the relationship between two routines and the current developer's version (patch 13) of Xfractint will not run other formulas. It should run yours, but even with that I would expect some sporadic crashes. There are several variables that need to be addressed in the ultimate solution; 1) DOS versus Linux, 2) Extraseg usage versus far memory usage, 3) The order in which the routines get called such that the correct variables are set and memory allocated before they are used. I need to think about this carefully, and sleep on it (several times), to ensure I get the solution right this time. For now, it is best to use patch 12. Jonathan
participants (5)
-
John Wilson -
Jonathan Osuch -
Niekamp -
Paul N. Lee -
Timothy Wegner