161) Message boards : Number crunching : New Version (Message 8419) Posted 15 Jul 2004 by Christopher Hauber Post: Just grabbed the new copy and at first glance I'm not seeing any visible changes. Anybody know what was changed? And I guess for whenever Rom or another developer reads this, it would be nice to have a version history available so people could see what was actually being changed/fixed. Chris

162) Message boards : Number crunching : just want to know ...... (Message 8416) Posted 15 Jul 2004 by Christopher Hauber Post: Looks ok. It should download automatically after it sits there for a minute or two based on your account settings. I just downloaded some workunits a few minutes ago and it worked fine so if it went down it just went down. I'm also still running 3.19 but I don't THINK that would cause a difference. Check your settings and let it sit a bit. If it still does nothing, do a manual update (projects tab, right click, select update now). Chris > do u people get WU ? > loaded setiboinc 3.20 and now i just get > SETI@home - 2004-07-15 21:42:16 - Sending request to scheduler: > http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi > SETI@home - 2004-07-15 21:42:23 - Scheduler RPC to > http://setiboincdata.ssl.berkeley.edu/sah_cgi/cgi succeeded > > Nothing further . is this right for now (project off?) or is there problems > with my settings ? > > >

163) Message boards : Number crunching : How to find rank if > 1000 (Message 8397) Posted 15 Jul 2004 by Christopher Hauber Post: Seems they fixed the bug. I imagine one of the XML stat files has the information but I don't know much about that. There are others around here who do though. > Is there any way to find your rank if it is greate than 1000? You used to be > able to just set offset=5000 if you wanted that but they seem to now enforce > the 1000 limit. > >

164) Message boards : Number crunching : HTTP_ActivePerl_Overflow attacks (Message 8358) Posted 15 Jul 2004 by Christopher Hauber Post: Drrr! That makes perfect sense. I'm surprised I didn't pick up on the cookie thing earlier. Although I don't know anything about goggle so I will just have to take your word for that but it makes sense especially since it came straight from EA. It's good that it is just a matter of the site mirroring, and kind of ironic that the images that WEREN'T being complained about actually caused it rather than the images what were. Hmmm... Just for the record, my last post in this thread is still one I stand by even though the pictures in question seem to have been cleared. > I don't think there's any relation to the name of the image (goggle~1) and the > goggle software. The image from Sim City actually exists on their server by > that name. The goggle software doesn't allow you to search for images > yourself, it runs on your own computer and just changes your desktop wallpaper > to a random image it found through google. Google allows you to search for > pictures, but goggle doesn't. > > The cookie is easy to explain. Remember that the forum here does not have any > actual pictures on it except the ones that it uses itself (the seti logo at > the top and the little heads beside usernames with profiles). Every other > picture you see is called from a remote location by your computer. The > forum only sends you the html code to allow your computer to go to the remote > site to fetch the image the code calls for. When you load any page with any > images at all, your computer is directly contacting the site that hosts that > image to get it. > > The Sim City site put a session cookie on my system when it fetched the image. > Nothing harmful in that. >

165) Message boards : Number crunching : HTTP_ActivePerl_Overflow attacks (Message 8351) Posted 15 Jul 2004 by Christopher Hauber Post: Well that is the most sound argument/reason that I have seen yet. I do understand really frequent posts like that could be a bit excessive and perhaps should be limited some. But I do think that periodic random posts are kind of fun. Shakes things up a bit. And things do need to be shaken sometimes. :) > It could have happened to anyone anytime anywhere. I still support the > pictures, > > ------------- > > I agree with those supporting pictures for they could be sometimes funny or > usefull in a technical discussion for example. But I do not agree posting > pictures just for posting pictures. It is pretty anoying and sometimes > irespecfull. The problem with Guido's behavior is that He join a discussion > with > a "and now for something completely different" message with a hughes pics of > bill gates. What the f..k does it bring to the discussion. To who he is > adressing, what is it he wanted to express, does it contain any personal > attack that only him can understand ? I have nothing personnal against him. > It's just that the very vast majority of people here have asked him politely > to stop doing > that but he just keep on posting more and more. He says he wants to improve > his English well that very good for him and discussion board is a good place > to practice, thats what I'm doing right now, but he wont improve anything by > posting pics. > > > >

166) Message boards : Number crunching : HTTP_ActivePerl_Overflow attacks (Message 8346) Posted 15 Jul 2004 by Christopher Hauber Post: It is interesting that the picture in question came from Electronic Arts. If that really is true, then I don't think EA would be very happy about that if they knew. The theory for the problem seems to be a decent theory, but it does not have anything to do with complaints for his pictures though. I don't think there was an mal intent, just an honest accident. Especially considering the server that the picture is located on. It could have happened to anyone anytime anywhere. I still support the pictures, but I would suggest images found to cause security problems be removed by their authors or the SETI staff. Chris > ----------------- > > The be honnest that was my first thaugh we the discussion point out that it > could be a somehow related to some images. If that is the case then it's a > strong argument in the balance for those complaining about Guido's pics. > > > > >

167) Message boards : Number crunching : Joining a team (Message 8339) Posted 15 Jul 2004 by Christopher Hauber Post: Thanks. Your English is better than putting German into Babelfish and having it translate :) > not in a team before : all the credits for the team > in a team before:only this credits from this time jo join the team > the old credits are in the team were you come from > sorry bad english i know > Ich höre immer gerne zu,wenn ich auch nicht immer belehrt werden möchte ;-) >

168) Message boards : Number crunching : Joining a team (Message 8324) Posted 15 Jul 2004 by Christopher Hauber Post: Just a quick simple question that I don't feel like going and finding the answer to. When you join a team, and have never been a member of a team before, do your existing credits get applied to the team you join or do only the credits that you earn during the actual time you are a member get applied?

169) Message boards : Number crunching : Cache sizes (Message 8310) Posted 15 Jul 2004 by Christopher Hauber Post: He said in another thread that a developer did try to fix some bug with certain clients or something to that effect. I don't remember exactly which thread it was, but if you click his name you can find all the posts hes written and find it pretty easily. Chris > Thanks for the update, Rom! Do you know why the times were hiked up in the > first place? > > ------------------------------------------------------ > We are Borgholio. You will be assimilated...bunghole! >

170) Message boards : Number crunching : New game folks: Guidos small picture thread not bigger then 10 kb (Message 8306) Posted 15 Jul 2004 by Christopher Hauber Post: Excuse me "Mr. Analog Modem"? Your signature is a bigger download than most of his pictures. Most people's signatures are bigger. Plus, the text of the web page downloads before the pictures, so you can still read. If you get charged for data transfered, then just turn the pictures off in your browser settings because you are going to have a lot more picture downloads on other web sites. If you want to complain about something, pick something worth complaining about and back it up with a valid statement. > Thinks witch are important to get little childs learn something. > > First and last point: ignore them! > > This means don't reply to his threads if thier conntent consists of pictures. > I have not something against pictures, but please Guido not in this ammount. > Here are people with an analog modem, so come on! > > The whole is more then the sum of its particles. Aristoteles > Best wishes from Berlin(Germany), where you can't see the Milky Way ;-)! > > Basti >

171) Message boards : Number crunching : Lurching from one farce to another... (Message 8019) Posted 14 Jul 2004 by Christopher Hauber Post: Well Rom, I'm not terribly surprised. Like I said before, all things considered, I think the dev team has done a pretty good job getting things fixed as they break and improving them so they are less likely to break again since the intitial release. There's still a lot to fix, but it seems like with the current configuration the server side has at least been stabilized. Now maybe more work can be done to fix and improve the client, like the proxy authentication problem ;) Chris > Just for the record, the estimates we had were botched because of how the > project came out into the public domain before it was officially launched. > > And before anybody says 'we should have controlled it better', I'll add that I > personally didn't want to screen every request for access to the system and > neither did any of the other Devs. 20,000+ people and growing. > > ----- Rom > BOINC Development Team, U.C. Berkeley > >

172) Message boards : Number crunching : Lurching from one farce to another... (Message 7984) Posted 14 Jul 2004 by Christopher Hauber Post: lol. I read your your threads about the pictures and all the "team talk" and found it pretty amusing.. but I've never really cared too much to join a team. Maybe I will eventually .. we shall see... nothing against the teams.. just assume be a loner for the time being.. The fresh start of BOINC does make joining a team a bit more appealing though even though I like being my "own" team :) > Chris, you are right. And incidentally, I noticed you're not a member of a > team. Care to join??? haha > >

173) Message boards : Number crunching : Lurching from one farce to another... (Message 7973) Posted 14 Jul 2004 by Christopher Hauber Post: Oh let him post his pictures. I've seen others who like it too. Personally it kind of makes me laugh a little because it provides a little randomness. Besides, it's already been pointed out that most of his pictures are smaller in size than most people's statistics signatures. He seems pretty easy going and good natured. Kind of a big picture sort of person. :) Chris > 1. the hole SETI community are asking you to stop posting pics. It has nothing > to do with nationality. You are getting on a very dangerous ground if you > start to threat "folks" from their nationalities. Big big mistake. > > 2. in your case it must have stop changing when you were 4 years old. > > 3. ? > > 4. I don't need you to tell me how to think. > > > > >

174) Message boards : Number crunching : Lurching from one farce to another... (Message 7966) Posted 14 Jul 2004 by Christopher Hauber Post: I didn't say impossible. I said difficult. Especially in a reasonable timeframe. Especially with limited funds. Especially with limited manpower. I still contend that for the type of project, it was the best way to really get a good feel for what needed to be done and fix it fast. Prediction models don't always work. Human response is extraordiarily hard to forsee. The addition of it being a wide area project (ie-internet/global) it's difficult to simulate that kind of pressure. Those simulations also don't necessarily push certain server applications to their limits and expose weaknesses in them. And incidently, having worked a total of more than a year and a half at grocery stores, those "predictions" aren't very reliable -- because no mathematical model known to man can predict human response. It can only give a vague idea. Seems to me they had a vague idea, but might not have thought to take certain things into consideration and I can't fault them for that. Those other projects are commercial with lots of money backing them. Airbus involved public safety, ergo very stringent checks. Big projects like communications satellites and such CAN afford to have bugs under the condition that they are minor and can be fixed without a lot of difficulty, and if major, can be fixed REALLY fast. But it's bad for them to have that kind of thing because they are trying to sell a service using them. No product=no money. Still though, bugs don't necessarily cause a product not to work. BOINC software works. The program BOINC uses to process SETI's workunits (as well as the one for Predictor) works. The servers work. The BOINC application and the servers have their problems, but the servers have really come a long way since it went public. But projects that are exploring or expanding relatively untouched areas, especially when scientific intensive, regardless of funding have problems. NASA has that happen all the time. The science behind these things isn't standard and isn't easy. Like Paul said, it's amazing it works at all. One other thing about this. I would venture to guess the folks at Berkeley/SETI are really interested in pushing forward with expanding the project to handle more specialized types of workunits (like broadband signals, 8-bit signals, data from other dishes, etc) and you can bet that before they do that, they want everything else to be in place operating smoothly on workunits that have already been processed and those that are not as vital before putting out a more advanced SETI "core" to process that other data. and if the rate of getting rid of bugs continues like it has the last few weeks, well then I should expect that things will be running smoothly enough to prepare those other services as well. You can think I'm wrong if you want. But it's what I have seen through helping Microsoft with its Internet Gaming Zone transition to the web, helping programmer friends test and improve some of their rather complex programs, and even some of my own programming experience. I would frequently run a program knowing I had errors just to see what else might pop up and other times when I thought it was done and had something else pop up. Really though, my gripe is for people like The Nameless Knight who started this thread who seem to complain about problems without even a hint of consideration for the complexities of releasing a computer program/project, especially one of this magnitude. Chris > I do agree that they're doing hard job and that SETI is not like a paid > service. > > I do not agree on the fact that it is impossible to predict the load or the > demand, call it the way you want on a system when you run it for the first > time. Being in engineering my self I have participate in a lot of commercial > project mostly in telecoms. I have seen many commercial lauch of satellite > communications projects like THURAYA or IRIDIUM not to include INMARSAT > improvement and satellite lauch. These kind of project simply can't affort > having bugs when lauching to public. > > What about airplanes ! Imagine the first commercial flight of an Airbus with > nothing onboard working. The company could not come and say well there was not > enaugh oxygen on board but we could not predict that our passengers would > breath so much. No it doesn't make sence. It is all a matter of will, > technical ressources and money. > > Imagine AT&T had just changed the new york area branch exchage (just > example) and the lines are on and off for a month ! Did it ever happened, I > don't think so. To make it simple, if we were all paying to run SETI and if > there was other project of the same kind in competition against SETI, the > whole new BOINC joined project would have worked 100% the first day. Period. > There are a lot of simple mathematical tools to predict the demand on a > system, just like grocery store can predict that between 9 and 10 they will > have a certain amount of customers, staying in for a certain amout of time, > line up to pay with a certain amount of items, and then decide how many > cashier to open. > > It's just as simple as that. > > Best regards > Marc > >

175) Message boards : Number crunching : New game folks: Guidos small picture thread not bigger then 10 kb (Message 7921) Posted 14 Jul 2004 by Christopher Hauber Post: Got it now.. I was close. Thanks :) Chris > or > i love to hear what the peoble say but i will not allways instructed > Ich höre immer gerne zu,wenn ich auch nicht immer belehrt werden möchte ;-) >

176) Message boards : Number crunching : New game folks: Guidos small picture thread not bigger then 10 kb (Message 7885) Posted 14 Jul 2004 by Christopher Hauber Post: Guido, I'm a bit rusty on my German and was just wondering if I was reading your signature right. I'm seeing something along the lines of: I always like to listen (enjoy listening), when I don't also want to be learning. I think I'm missing something here because I'm not quite sure I understand that.. Anyway.. just curious :) Chris > this small enough > <a> [/url] > Wireless Signals of the Past >

177) Message boards : Number crunching : Lurching from one farce to another... (Message 7873) Posted 14 Jul 2004 by Christopher Hauber Post: I don't know how "bad" beta was because I never used it because I wanted to see my progress and guage how much work I was doing. As "many" people as there were during the beta, there were MANY more that switched over when BOINC was officially released. That means in just a couple of days you have not only those beta users updating to it, you also have many times that switching over as well. You go from a fairly easy to handle load to an extremely high load potentially causing an overload as apparently happened. Also, the stuff happens so fast that and all at once that problems you had no idea were there will surface, and problems that under normal conditions aren't really problems at all becoming detrimental to maintaining system integrity. I don't know how much work SETI had stored when they released BOINC, but it very well could have been close to 2 weeks. I am guessing that people switched to BOINC a little faster than expected. Couple that with the fact that now essentially every client was caching days worth of workunits, that supposed 2 weeks could be gone in just a day or two, which it was. And I think they did a fairly decent job of rectifying the situation when they realized how fast we drained the stockpile considering budgetary and time constraints alongside the enormity of the project. There wasn't a very good way for them to tell that their servers were going to be that ill-equipped to handle the load and the best way to find out is also the worst way and thats what they did. There are lots of problems that beta testing just doesn't find very well or takes too long to find. Releasing into the public finds those problems very quickly. SETI can get away with that because they aren't charging money for it. They are non-profit. Profit companies can't take advantage of that tool and are forced to find more of the bugs before release. Microsoft is the perfect example. They have tons of money and people. They charge a bundle for their software, and then when you get it, its full of bugs that they just couldn't find because of being behind a huge corporate firewall and the exhorbanent amount of code and just plain time. But people still PAY for it anyway. Take SETI now, nonprofit with a limited staff (and just because they employ a lot of students does NOT make them any less worthy as programmers.. they may be better than a lot of companies) and a massive project that dwarfs anything out there. What do you expect? Updates are far more frequent than SETI clasic, until today there has been little shortage of units since the first big one, the software isn't perfect but pretty stable. Yea there are a lot of problems and they are a little frustrating, but good golly I applaud the people at Berkeley for their work on this. They seem to have been working all night all day and all weekend for a while now getting things fixed, arranged more efficiently, troubleshooting, etc. I would browse around at all times of day and week and find new messages from SETI staff regarding problems and fixes and such at times that you would least expect there to be qualified personel on staff to fix problems. I'm not saying don't complain or mention problems when you find them. They need to know and it can help other users, but don't complain about how horrible it is and how much it sucks and that it is a piece of junk. This is just how things work for a large scale cooridinated computer project works for getting off the ground without having to wait forever and a day. I like it that way because I know the problems are more likely to get fixed promptly if they are found now than later. Give it a chance, don't scare off new users who want to help but are nervous about how it works or putting some group project software on their computer. I think a lot of people forget that this is essentially 2 experiments rolled into one. Maybe even three. The distributed computing project is an experiment in itself dispite being around for a while. BOINC is an extension of that to try providing a platform for all kinds of projects to use distributed computing. And then there is the obvious one, to find some possible sign of life out there for believers and nonbelievers alike. Anyone who has ever done a science experiment should know that they don't always work out the way you plan and you can't always think of half the things that happen. I've been very getting very agrivated by people trashing BOINC and it's developers when they apparently have no understanding of things. I am not a programmer, but I HAVE programmed in the past and as minor as those projects were in the grand sceme, it helped me to understand things a lot. I've also worked in a beta "transitional" phase before (Microsoft's Internet Gaming Zone when they went from client based to web based) and they had a lot of the same kinds of problems. Bring stuff up to see if it works and have it crash, fix the problem and try again until it worked. And much of that was AFTER they publicly released the web version AND shut down the client. At least we still have the original client for the time being. I'm sorry for the length of this, I just had to say my piece. Please though, when you have complaints, just think about what it is you are complaining about. If it's worth complaining about, at least try to understand what may be behind the problem instead of ripping into it. I'm done. I have work to do. Happy crunching to those that are easy-going about the problems and try to help others with their problems when they can. Chris

178) Questions and Answers : Wish list : Retaining Transfer log (Message 7521) Posted 14 Jul 2004 by Christopher Hauber Post: Not everything done is in the log. The only time that I know of for sure that it doesn't display uploads is when I've told it to update rather than it handling updates automatically. The completed units simply disappear. I've seen many times in the log where it says something has been uploaded. But the whole point to this was not that it does or doesn't show it. It was WHERE it shows it. I just think that the transfer tab should have all current, pending, and past transfers with all relavent information in a sort of table form. Right now, for me and probably most people with broadband and maybe even dialup, the transfer tab is pointless. And in my opinion the main log is cluttered and I think shifting nitty gritty specifics of the transfer information would help free the log window of a lot of that clutter. Streamline the main log, retain certain information in a clean, readable format in another place where it seems better suited anyway. > > I know they are (although it rarely seems to report uploaded units in the > log) > > Everything it does is reported in the log. Uploads look like this: > > SETI@home - 2004-07-13 16:55:31 - Started upload of > 04ja04aa.7402.13138.386086.218_1_0 > SETI@home - 2004-07-13 16:55:41 - Finished upload of > 04ja04aa.7402.13138.386086.235_1_0 > > If you aren't seeing them, try saving the entire log to a text editor, and > search for "upload". > > > <a> [/url] >

179) Questions and Answers : Wish list : Retaining Transfer log (Message 7370) Posted 13 Jul 2004 by Christopher Hauber Post: I know they are (although it rarely seems to report uploaded units in the log) But the way its done looks horrible. I want to see transfer messages in the transfer log, NOT the main message log. I don't care if it says that something IS downloaded or IS uploaded along telling that it was successful or had some error, but that is all. I don't want to see the specific details and numbers in the main log. Get rid of filenames, workunit names, throughput, and other specific details and put them in the transfer section as a transfer log/history. I just would like to see the main message log cleaned up a little and made to be easier to read. Right now it's a pain to find something specific. > All the relevant transfer details will be in the message log. > > <a> [/url] >

180) Questions and Answers : Wish list : Retaining Transfer log (Message 6981) Posted 12 Jul 2004 by Christopher Hauber Post: I would like to all file transfer details (ie-workunit name, file name, troughput, current transfer speed if applicable, file size, etc) to be moved to the transfer tab and retained there even after completion of the transfer. I think the only thing to put in the general log would be that a file or workunit is being started and/or finished with or without an error, and what the file is (workunit, core program, SETI Processor, project logo, whataver so long as it is something that means a little more than just the file name). I just don't think that it should be purely for active transfers. I never see any transfers because they either never show up in the transfer log (I've never seen a workunit upload show up) or they are completed so quickly that they are gone before I have the chance to look. Plus I think it would clean up the general log a little and make it a little more understandable, and it would make file transfer details easily accessible and readable. Chris

Previous 20 · Next 20

 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.