.... restarted now (some minutes ago) and now I have full work buffer.

End of panic ON?

I Hope.

Franz

I'm sure that BOINC and the projects depend on vast and unpaid efforts of enthousiastic volunteer tetsters and pro's sacrifying their spare time. I just want to stress that very much could be gained by putting more attention to the interface, and make it much more "seamless" and tansparent. That is not a matter of perfect (bug-free) ) code, I think, more of an understanding how "ordinary" people — used to simplicity and selfevidence when dealing with computers and not prepared to spend hour's browsing the Message Boards (and Wikipedia) — may be frustrated, lose interest and eventually skip the whole enterprise.

I'm not so certain it's a matter of the developers not knowing or understanding how "ordinary" people use their computers and their expectations that go along with it. In fact, I think they're very aware and that's why BOINC development continues, and why feedback is very important, not sordid complaints such as "I want the old SETI back" or "BOINC sucks" without specific detail and without attempting to make it work or giving feedback on how to make it easier.

And that we don't want do we? Of course we like everybody to join and in that case you can speak of crunching-power! Maybe something is happening after all. GridRepublic seems promising, not sharp yet and not for Mac I'm afraid.

I absolutely agree.

I'm sure that BOINC and the projects depend on vast and unpaid efforts of enthousiastic volunteer tetsters and pro's sacrifying their spare time. I just want to stress that very much could be gained by putting more attention to the interface, and make it much more "seamless" and tansparent. That is not a matter of perfect (bug-free) ) code, I think, more of an understanding how "ordinary" people — used to simplicity and selfevidence when dealing with computers and not prepared to spend hour's browsing the Message Boards (and Wikipedia) — may be frustrated, lose interest and eventually skip the whole enterprise.

I'm not so certain it's a matter of the developers not knowing or understanding how "ordinary" people use their computers and their expectations that go along with it. In fact, I think they're very aware and that's why BOINC development continues, and why feedback is very important, not sordid complaints such as "I want the old SETI back" or "BOINC sucks" without specific detail and without attempting to make it work or giving feedback on how to make it easier.

And that we don't want do we? Of course we like everybody to join and in that case you can speak of crunching-power! Maybe something is happening after all. GridRepublic seems promising, not sharp yet and not for Mac I'm afraid.

I absolutely agree.

There's a lot of people out there who'd sure like to put their computers to use for the benefit of the projects that has joined BOINC for that purpose. And of course the removing-procedure cited above is far beyond their scope of skill/experience/interest - and why not. It should be a problem for Seti, Einstein and the other projects to convince BOINC about the need for an even more seamless installation/uninstallation procedure, don't you think? BOINC has more to accomplish before you can speek of perfect management in this area.

No one said the BOINC Mac code was even close to perfect. Please remember that most of the Mac code is done by volunteers, as most of the SETI staff have Sun (Solaris)/*nix/Windows machines to work on and use for testing. Since most users use Windows, that is where most of their development time is spent perfecting.

I think it is time to summary the list of the available ports :

<snip>

Alex Kan+Rick Berry ; http://writhe.org.uk/seti@home
- MacOS X optimized for G4 and G5

It's worth noting that the site hosting the Mac clients hasn't been updated since December 13, 2005 (I've been out of touch with Rick for a while), and therefore don't have any of the Enhanced Mac clients that I've released in the past few months. I've been hosting my clients at http://tbp.berkeley.edu/~alexkan/seti/, but haven't gone through the trouble of putting a proper site up. There'll be one up there in the future, I promise!

— By the way, these wu's waiting for upload are occupying RAM are they not? Am I better to suspend S@h alltogether at the moment and would that clear the RAM from s@h-wu's waiting for upload?

WUs that have been completed are NOT stored in RAM, as this would cause them to be lost if you turn off your machine!!!

While WUs are being crunched, they are held in RAM, as well as the original which is kept on your hard drive. By default, the results crunched are svaed to HDD every 60 seconds, so if you stop BOINC, turn off your machine, have a power failure, etc, the work is not lost, and resumes when you restart BOINC.

Similarly, the result is stored on HDD until you upload to S@h.

In a result you finished:
OS X optimized S@H Enhanced application by Alex Kan
Version info: OS X Altivec (G4/G5, G5-optimized v6-nographics) V5.13 by Alex Kan

That shows you are using an optimized application.

If you look at your results page, when you made the change over, your speeds have more than halved, and up to 1/3 less time than before.

Where do get the SETI 5.17 ? - 5.17 is only for testing in the BETA project, and not for use in the main project.
If you wan't to use the optimized app by Alex Kan, you need to stop the boinc client, and put the app_info.xml and the binary into the projects/setiathome.berkeley.edu folder. But attention, if you place the app_info.xml, the boinc client send back all the WUs in the cache and the WUs in progress.

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>seti_enhanced-ppc-v6</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>513</version_num>
<file_ref>
<file_name>seti_enhanced-ppc-v6</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>517</version_num>
<file_ref>
<file_name>seti_enhanced-ppc-v6</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>

[…]

Fri Sep 1 20:39:05 2006|SETI@home|State file error: no application version setiathome_enhanced 517

So, what could one say? :-(

Your edited file looks just fine to me—exactly like mine, at least. Are you sure it was saved as plain text (without formatting, and not as RTF or similar), replacing the previous version? Using the same character-encoding as the original? If so, I’m at a loss to explain why it wouldn’t work.

If you open it with a browser (or some other application that parses or verifies XML), does it read “setiathome_enhanced seti_enhanced-ppc-v6 setiathome_enhanced 513 seti_enhanced-ppc-v6 setiathome_enhanced 517 seti_enhanced-ppc-v6” with nothing missing and nothing extraneous? If so, again the file itself would seem fine, implying that BOINC is unable to find or read it for some reason. All I can think of to suggest at the moment is to start over: remove the app_info file (with BOINC stopped), let BOINC download the standard app, verify that it’s back to normal—slow & steady—then have another try with Alex Kan’s app, this time using the new version of the app_info file to start with.

I’m going on vacation for a couple of weeks; I hope someone else can step in with some better advice. Sorry, and good luck!

Hi, whoever replaces Odysseus who has gone fishing. I checked the app_info.xml file shipped with Alex' ppc_v6 and which I had edited according to Odysseus suggestions, by opening it in my browser Safari, and the string seemed to match exactly the one Odysseus shows in his reply nr 413345. So I went on, quitted BOINC, removed the file from the s@h folder, restarted BOINC after having reinstalled the app from my backed up package. What happend after that is not quite retractable I'm afraid. I never tried to reinstall the ppc_v6 since the app seemed still to be residing in the s@h folder. Though Mac seemed to happily crunch both s@hWUs and s@hβWUs at the moment, in both cases using 517 that resided only in the s@hβ folder, there were some problems left wich I cannot remember now. As I would like to make a new install of ppc_v6 in s@h (and hopefully later in s@hβ) with the edited app_info.xml there from the beginning, as suggested by Odysseus, I was not saisfied with having the ppc-6 client left there. So I decided to remove the app and once more reinstall BOINC from my back-up package hoping the original client should now be installed in the s@h folder (as it should have allready been according to Odysseus) and it was, sort of, but another version 513 (?) of set_enhanced. I don't remember if 513 was the one originally shipped with s@h, before my "upgrading" to ppc_v6, or if it came when I attached to
s@hβ and the problems began. Anyhow both 513 and 517 figures in the ppc_v6's app_info.xml file. At the moment 513 governs Mac's crunching s@hWUs and 517 (as before) the s@hβWUs. If I look at the graphics I see some differences between 517 and 513 and the graphic showing the 513 grunching has features that I don't remember to have seen before. So the 513 is a new (but older?) player I guess.

Leaving the question of once more trying to "upgrade" to pp_v6, I would like to have the 517 client working with s@h as well, since it seems to do the job i another and (presumably) better way. How do I accomplish that? If I were not afraid of loosing my credits I'd consider start all over again. Deinstall BOINC etc.

By the way, how come that my signature BOINCstats has not uptaded since 29th?

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>seti_enhanced-ppc-v6</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>513</version_num>
<file_ref>
<file_name>seti_enhanced-ppc-v6</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>517</version_num>
<file_ref>
<file_name>seti_enhanced-ppc-v6</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>

[…]

Fri Sep 1 20:39:05 2006|SETI@home|State file error: no application version setiathome_enhanced 517

So, what could one say? :-(

Your edited file looks just fine to me—exactly like mine, at least. Are you sure it was saved as plain text (without formatting, and not as RTF or similar), replacing the previous version? Using the same character-encoding as the original? If so, I’m at a loss to explain why it wouldn’t work.

If you open it with a browser (or some other application that parses or verifies XML), does it read “setiathome_enhanced seti_enhanced-ppc-v6 setiathome_enhanced 513 seti_enhanced-ppc-v6 setiathome_enhanced 517 seti_enhanced-ppc-v6” with nothing missing and nothing extraneous? If so, again the file itself would seem fine, implying that BOINC is unable to find or read it for some reason. All I can think of to suggest at the moment is to start over: remove the app_info file (with BOINC stopped), let BOINC download the standard app, verify that it’s back to normal—slow & steady—then have another try with Alex Kan’s app, this time using the new version of the app_info file to start with.

I’m going on vacation for a couple of weeks; I hope someone else can step in with some better advice. Sorry, and good luck!

I've got a new problem I'm afraid. Following a suggestion of your's I attached to another project — s@hβ —by the command in BOINCs tools-menu and from then on no new s@hWUS were accepted to be delivered. Every request for new tasks was denied […]

Thu Aug 31 22:08:39 2006|SETI@home|State file error: no application version setiathome_enhanced 517

You just had to pick that one, didn’t you! ;)

There’s a known bug (which I gather has been fixed in an alpha version of BOINC, yet to be officially released) where BOINC gets confused by having two project apps with the same name but different version numbers, despite the fact that they belong to different projects and reside in different folders. There are a few threads about it in the Number Crunching forums, here and at S@h Beta/Astropulse. The simplest solution is to edit the app_info.xml file that came with Alex Kan’s cruncher; if you make it look exactly like the one I posted in Beta NC you should be back in business. Stop BOINC, edit the file to add the <app_version> section for v5.17 (using TextEdit or similar, saving it as plain text), and relaunch BOINC. Your optimized v5.13 results here will be shown in BOINC Manager as v5.17 and reported that way as well, but the output file will continue to identify them correctly as coming from Alex’s app.

Note that while the above errors were happening, the ‘rejected’ WUs will have been discarded, but they’ll still show up in your Results pages as being in progress. They’ll go away eventually, after their deadlines have passed and other hosts have processed them. When the problem hit my G4 at work (when Beta v5.17 first came out) it was a long weekend here, so I was confronted with a daunting list of failures the next Tuesday morning … I’m still seeing more than a dozen of them, but they’re due to expire on Sept. 2 & 3, so I shouldn’t have to look at them much longer.