I am not very familar with this project and looked at the FAQ (for a while anyway) but didn\'t see this explained.

regarding my result:
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6191095

It shows a bit over 29 hours cpu cycle time and a successful conclusion (but obviously not a end of model)... but there is no credit claimed or granted. Did I (or my machine) do something wrong?

You will see I have other models now running for months and would hope they give me and my team some credits (they are!!!)

I am just seeing what is funny about this result.
(I ahd an other zero one but it had errors)

Thanks

The \"Over Success Done\" displayed on result pages is often misleading.
I don\'t know why it happens, but it does.
In the case of your model, it\'s crashed.

And you won\'t get any credit because that model never returned any trickles. These trickles contain science data, which is what you get granted credit for.

---

Backups: Here

Another item of BoincSpeak I don\'t like is the use of the word \'results\' for workunits we\'re running or have run, even before any results have been produced for the researchers.

Jumbo, if you follow the links in Les\'s signature to the project READMEs, you\'ll find much more detailed info than in the FAQs. I\'d particularly recommend the items on how to make boinc folder backups (so that if a model crashes it can be restored and continued), and in the README about avoiding model crashes, item #5 by Mike on avoiding the most common pitfalls.

---

Cpdn news

Another item of BoincSpeak I don\'t like is the use of the word \'results\' for workunits we\'re running or have run, even before any results have been produced for the researchers.

We\'re not running WUs, but \'tasks\' in the manager .....:o)

There are also a lot of crashed wus on SAP that are marked as \'success\' and the credits are in \'pending\' state: The question was asked to me twice last week, on the forum of team l\'AF.
As the Wus are in \'pending\' state and they worked several hours on them, the users expect some credits....

---

Arnaud

Hi Arnaud

Yes, tasks in the boinc manager but results on members\' server web pages. But tasks I think only refers to current crunching workunits whereas results are the complete historical list of a member\'s downloaded WUs.

If you were taking an examination eg in Maths you couldn\'t say that it was a result while you were still trying to answer the questions. Or if you were trying to to cook meringues, you couldn\'t call them a result while you were still beating the egg whites. The word result should be reserved to describe the outcome or the outcome so far.

Certainly, to use the word Success for a WU that has failed is crazy.

Les has started discussing this in the moderators\' section of the independent forum. Assuming that boinc imposes this vocabulary and not cpdn, we should think about changes we\'d like for boinc version 6.

---

Cpdn news

For example, what boinc manager calls the task name for a WU I\'m running is on my server web page called result name

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6740728

On that same server web page I also think it\'s unhelpful for newbies that the phrases host ID and computer ID are used to mean (I think) the same thing. When I started cpdn I had no idea what a host was, apart from a person inviting guests to dinner or a party.

---

Cpdn news

Hi Maureen,

In fact \'task\' was chosen by Dr Anderson to replace \'work\' in the Manager.
IMHO, \'work\', \'task\', and \'result\' should all be replaced by \'unit\' or \'work unit\': simplification of the vocabulary would be better for the \'newbies\'.

All that confusion is the result of the historic evolution of Boinc, the Manager and server side are probably not coded by the same people at the same moment.

AFAIK, Computer ID and Host ID are not the same thing: one computer can have several host ID. Just detach and attach again several times with the same computer and you\'ll see the host ID change: don\'t know what\'s the point of such a mechanism...(except that you can download new work when you have passed the quota of WUs)

Your exemple is good: \'Result\' is used on the website, but the state of the result is unknown and the server state says \'in progress\': can a result be in progress and unknow....:o)

---

Arnaud

A football match can be \'in progress\', so that the \'final result\' is unknown, but the \'result so far\' can be known.

Beginners should just learn the words, the same as in any endeavour.
(Golf comes to mind.)

I agree to some extent, Les, but I don\'t think a sporting nation like Australia would call the actual playing of a match a \'result\'. And especially not before the match has even started. You couldn\'t say that next month a rugby result will be played between England and Australia.

Boinc does need to distinguish between the workunit itself, of which several copies may have been created to run on different machines, and the single copy of it that a member is crunching. On the forums people usually say \'workunit\' for both meanings. It might be better to say \'workunit\' for the first meaning and \'task\' for the second. Then keep the word \'result\' for what has happened to it so far.

---

Cpdn news

The word result IS used \'down here\', once the game has started.

But all of this is getting away from the point that I made about the original post:
The word \"Success\' is completely incorrect to describe a model that has failed.
A more correct word would be \"Failure\".

Addendum
The word Success may have been used in this case because it was one where BOINC thought that the model had completed, and displayed, for a while, 100% under Progress. The error 161 with all of the zip file names would support this idea of it thinking that the model had completed, where as it hadn\'t even reached the first trickle.
As for what caused:

<![CDATA[
<stderr_txt>
(null): cannot open input file dataout/atmos_restart.day
(null): cannot open input file dataout/ocean_restart.day

I don\'t know. Some other program using the hard disk intensively at the time?
A virus checker locking files?

But it also says:

Fatal crash! :-(

I can understand the use of task.

On some projects like CPDN you work on a \"Work Unit\", on other projects, like SETI, you work on \"Results\" of a \"Work Unit\". Since you are not always working directly on a \"Work Unit\" in all projects, he used a middle term of Tasks.

Plus I am not a person of many languages, but BOINC is, and maybe Tasks is a better work that translates into more languages than other words previously used?

Just my thoughts on the word Tasks.

---

One of Alan Thurgood\'s crashed models also shows OVER-SUCCESS-DONE:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6169968

\'Done\' means that \'no error occurred\' and it\'s used when a model that was advancing correctly is aborted. But this model wasn\'t aborted - it crashed with a -161 error code. I wonder whether there are cases where the server doesn\'t detect the error code?

---

Cpdn news

I had an error-code-0 which also showed over/success/done. That particular example crashed when I tried to shut it down with task manager (I know better now :-) )

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ