| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Same CPID for 2 different accounts |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 60
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
No.
|
||
|
|
Saenger
Advanced Cruncher Germany Joined: Dec 28, 2005 Post Count: 68 Status: Offline Project Badges: 
|
No. Can you at least tell me where the user.gz or user.xml is located on this server? I could look up whether both numbers are included somewhere. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
|
||
|
|
BobCat13
Senior Cruncher Joined: Oct 29, 2005 Post Count: 295 Status: Offline Project Badges: 
|
TSWB Stats 'N Stones shows both accounts under the CPID, although they are not combined.
http://statsnstones.tswb.org/userdetail.aspx?...eb95d0492e94ca15007c2df40 Using the same CPID at boincstats doesn't show both accounts. It would appear to be something Willy needs to correct for his site to show both. http://boincstats.com/search/all_projects.php...eb95d0492e94ca15007c2df40 |
||
|
|
Saenger
Advanced Cruncher Germany Joined: Dec 28, 2005 Post Count: 68 Status: Offline Project Badges:
|
TSWB Stats 'N Stones shows both accounts under the CPID, although they are not combined. http://statsnstones.tswb.org/userdetail.aspx?...eb95d0492e94ca15007c2df40 Using the same CPID at boincstats doesn't show both accounts. It would appear to be something Willy needs to correct for his site to show both. http://boincstats.com/search/all_projects.php...eb95d0492e94ca15007c2df40 It's something Willy needs to be made aware of, as this is not the expected behaviour of a BOINC project to split CPIDs over several entries in the exported XML. One man, one vote, and one CPID, one XML-entry. The possibility of two entries in the XML (and I found them as I looked for them, quite far apart from each other) is nothing that anyone would ever have thought about, it should be impossible as it is severely wrong. |
||
|
|
Saenger
Advanced Cruncher Germany Joined: Dec 28, 2005 Post Count: 68 Status: Offline Project Badges:
|
Off-Topic: How long is Edit possible? Edit: Dunno what happened, but this
---------------------------------------- was missing as I wanted to edit the former postOn topic: I found one stats site that includes both: James' netsoft Edit: And I got an answer from Willy: It's simple: the CPID is supposed to be unique. When the XML file is imported and there are duplicate CPID's then only the first one will be imported, the next one(s) can't be imported because they violate the unique key. ---------------------------------------- [Edit 2 times, last edit by Saenger at Jun 5, 2009 8:13:21 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Willy is wrong.
I think he has been told this before. Using CPID as a primary key is woefully inadequate. |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
Willy is wrong. I think he has been told this before. Using CPID as a primary key is woefully inadequate. As a stats-site, using the CPID is the only thing he can do, since user-names aren't unique neither within BOINC-projects nor between BOINC-projects, and of course the email-address that is unique is for security-reasons not exported in the stats. As long as a user uses the same email-address, and is attached in all projects on a single host or has another route between hosts so CPID can synchronize, the CPID is the unique identifier, since the email-address is unique in BOINC, except for WCG that has a broken implementation that doesn't enforce unique email-addresses as they must do to not screw-up for BOINC-users, stats-sites and possibly also in other ways.  "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Ingleside, the CPID isn't even guaranteed to remain the same. Using it as a primary key is a huge mistake.
----------------------------------------The WCG implementation is not broken, BOINC specifies NONE of this stuff. This is a major weakness of BOINC, of course. Few things are properly documented. Using the email address as a key is a relic from early versions of BOINC. It was disastrous then, and it is obsolete now. Allowing duplicate member names was a particularly poor move - now there are literally dozens of people with a name of "Administrator", "Admin", and "root". BOINC failed in a major way here, so your defence of it is.... futile at best. edit: And I should point out that the stats export *contains* a primary key, guaranteed to remain the same for a specific account at that project. [Edit 1 times, last edit by Former Member at Jun 6, 2009 10:11:13 AM] |
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
Ingleside, the CPID isn't even guaranteed to remain the same. Using it as a primary key is a huge mistake. The changing nature has for years been a problem, especially as BOINC-projects stops running, but it is still the only thing stats-sites can use to generate cross-project-stats. The WCG implementation is not broken, BOINC specifies NONE of this stuff. BOINC-projects is expected to download the BOINC-source-code, or a virtual machine with everything working, and use this as a basis then creating their project. Even if a project wants to customize things, they'll still need to look on the supplied code to make sure it's working alongside the other BOINC-code. Whoever made the customizations for WCG clearly forgot to include the neccessary line about unique email-address, that showed-up in my code-snippet. This is a major weakness of BOINC, of course. Few things are properly documented. No arguments that BOINC has major weaknesses in the documentation-department. Because of this weakness, any project that makes any customizations must look on the code they're customizations, to make sure they don't forget something or screws-up something in the process. Using the email address as a key is a relic from early versions of BOINC. It was disastrous then, and it is obsolete now. Hmm, how long is it since you attached to a BOINC-project appart for WCG? Chooses "Attach to project...", selects any of the BOINC-projects from the list... Regardless of chooses "new user" or "exsisting user", BOINC-client asks for your... email-address... In BOINC, the email-address is the "key" that shows which account it is, just like it was the only "key" used back in SETI@home "classic". The only change is that under BOINC you'll also need to type-in a password... Allowing duplicate member names was a particularly poor move - now there are literally dozens of people with a name of "Administrator", "Admin", and "root". BOINC failed in a major way here, so your defence of it is.... futile at best. This has historical reasons, since BOINC was developed as the upgrade for SETI@home "classic", but with multi-project-support, they didn't want to force the SETI@home-users to change to an unique name, after having used their non-unique SETI@home-name for some years. This doesn't stop other BOINC-projects for enforcing unique user-name if they so wants to do it.... BTW, in the real world, very many people has the exact same name as someone else, and BOINC doesn't disallow people from continuing to using their name if they signs-up for one or more BOINC-project. But, even many people has the same name, most people has an unique email-address, either through work, school or whatever, and they can use this email-address to identify themselves uniquely... edit: And I should point out that the stats export *contains* a primary key, guaranteed to remain the same for a specific account at that project. You can't generate any cross-project-stats based on this key... "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
|