I only care about credits because they show me that my computers are continuously turning in valid work. However, I do not like doing work units for no reason. That's a waste of computer time.

https://boinc.bakerlab.org/rosetta/workunit.php?wuid=37380301

So, if I'm reading this correctly, the first user had an error. The second user didn't turn it in on time so my computer got the unit. However, the second computer turned it in before I did, and so when I turned mine in a day after I got it I get 0 credits?

That's a waste of 4 hours of computing time! And I don't like crunching work for no reason if the guy who turns it in late still gets credit when I turn it in almost immediately and I get 0 credit! That 4 hours could have been used for a work unit that actually needed crunching.

---

I agree with Matt. It is particularly unfortunate as the task that has been zero credited actually delivered more decoys in this example.

The problem is with the max values set for the workunit - 1, 2, 1

This means it takes a max of 1 error result, 2 total, or 1 success. In the usual case these limits all hit at the same time, when there is more than 1 error the WU stops replicating. Here there is an extra success result because a task that timed out did eventually deliver.

The max of 1 error is probably right, as it stops rogue wu from replicating too far (who can forget Xmas 2005?)

I'd suggest a higher value for the numbers of total and success tasks, keeping 1 as the error limit. Something like 1, 4, 4 would make sense in my opinion.

Secondly, as Rosetta advertises that it allows work back after the deadline, I wonder if it is possible to tweak the scheduler so that it either wait a lot longer than the deadline before sending out a replacement task, or to not send one out at all?

I hope credit will be awarded manually for this task and any similar - I know this has been done in the past.

River~~

I only care about credits because they show me that my computers are continuously turning in valid work. However, I do not like doing work units for no reason. That's a waste of computer time.

https://boinc.bakerlab.org/rosetta/workunit.php?wuid=37380301

So, if I'm reading this correctly, the first user had an error. The second user didn't turn it in on time so my computer got the unit. However, the second computer turned it in before I did, and so when I turned mine in a day after I got it I get 0 credits?

That's a waste of 4 hours of computing time! And I don't like crunching work for no reason if the guy who turns it in late still gets credit when I turn it in almost immediately and I get 0 credit! That 4 hours could have been used for a work unit that actually needed crunching.

Your computer claimed 50 credits and was granted 20. So what are you talking about?

Check Again Matt

43616155 302987 24 Oct 2006 0:11:43 UTC 25 Oct 2006 20:56:35 UTC Over Success Done 14,328.41 50.01 20.00

You are the owner of record for computer 302987

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

I only care about credits because they show me that my computers are continuously turning in valid work. However, I do not like doing work units for no reason. That's a waste of computer time.

https://boinc.bakerlab.org/rosetta/workunit.php?wuid=37380301

So, if I'm reading this correctly, the first user had an error. The second user didn't turn it in on time so my computer got the unit. However, the second computer turned it in before I did, and so when I turned mine in a day after I got it I get 0 credits?

That's a waste of 4 hours of computing time! And I don't like crunching work for no reason if the guy who turns it in late still gets credit when I turn it in almost immediately and I get 0 credit! That 4 hours could have been used for a work unit that actually needed crunching.

Your computer claimed 50 credits and was granted 20. So what are you talking about?

Check Again Matt

Maybe it changed between when I originally posted and when you saw it, genius.

I only care about credits because they show me that my computers are continuously turning in valid work. However, I do not like doing work units for no reason. That's a waste of computer time.

https://boinc.bakerlab.org/rosetta/workunit.php?wuid=37380301

So, if I'm reading this correctly, the first user had an error. The second user didn't turn it in on time so my computer got the unit. However, the second computer turned it in before I did, and so when I turned mine in a day after I got it I get 0 credits?

That's a waste of 4 hours of computing time! And I don't like crunching work for no reason if the guy who turns it in late still gets credit when I turn it in almost immediately and I get 0 credit! That 4 hours could have been used for a work unit that actually needed crunching.

Your computer claimed 50 credits and was granted 20. So what are you talking about?

Check Again Matt

Maybe it changed between when I originally posted and when you saw it, genius.

So , before you bellyache , you should recheck.

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato

I only care about credits because they show me that my computers are continuously turning in valid work. However, I do not like doing work units for no reason. That's a waste of computer time.

https://boinc.bakerlab.org/rosetta/workunit.php?wuid=37380301

So, if I'm reading this correctly, the first user had an error. The second user didn't turn it in on time so my computer got the unit. However, the second computer turned it in before I did, and so when I turned mine in a day after I got it I get 0 credits?

That's a waste of 4 hours of computing time! And I don't like crunching work for no reason if the guy who turns it in late still gets credit when I turn it in almost immediately and I get 0 credit! That 4 hours could have been used for a work unit that actually needed crunching.

Your computer claimed 50 credits and was granted 20. So what are you talking about?

Check Again Matt

Maybe it changed between when I originally posted and when you saw it, genius.

So , before you bellyache , you should recheck.

How about you stop being a moron?

When I posted it had 0 credit. It was granted credit after that.

This seems to be too complex of an idea for you.

Go start trouble somewhere else.

For the validate section of each of those..
#1 says "invalid"
#2 says "valid"
#3 (Matt's) says "workunit error - check skipped."
is that the normal message for a WU turned in after a valid WU?

---

I have no idea.

The 20 credits are nice, but I started the thread to let the scientists know that it's annoying (and wasteful) to send out a work unit for a second time and then deny that computer's result when the late person returns his late.

I also got one of those "check skipped" units and got 20 credits after a delay. You usually have to check the actual WU result to see it. I have always gotten credit for failed units as long as they did some useful work. One of the things I like about Rosetta.

Bill

I'll look into possibly modifying the validator to fix this.

For the validate section of each of those..
#1 says "invalid"
#2 says "valid"
#3 (Matt's) says "workunit error - check skipped."
is that the normal message for a WU turned in after a valid WU?

First, to avoid confusion can we be careful with the terminolgy please - these three tasks are all the same WU. The name "work unit" covers the group of tasks sent out with identical data & identical seeds.

It is a task when it is sent out and it is a different task if it is sent again to a different host. It is a different task on each host, it is still the same WU on both/all hosts.

If I understand the jargon a task becomes a result if and only if it is returned to the server.

Understanding the jargon helps us to understand the error message: we deduce that the message "work unit error" does not refer specifically to Matt's result, but to the effect that result has had on the entire workunit by its very existence. At the moment of return the work unit goes into error even before the content of Matt's result is looked at.

By using the words "task" and "WU" interchangeably, we miss the subtleties of this kind of message. That is why I hope my pedantry over the terminology is useful.

If the first failed task had not existed, so that there were just two good tasks in that WU, Matt's result would have been validated. Probably (but we can't be sure) it would have been accepted and he would have got the credit automatically.

Hence my suggestion to use maxima of 1, 4, 4 instead of 1, 2, 1 when setting up future WU types. Then in cases like this where late return causes more than one task to be sent out, the work unit would not be in error and the component result would be analysed.

HTH
River~~

---

... I don't like crunching work for no reason if the guy who turns it in late still gets credit when I turn it in almost immediately and I get 0 credit! ...

Your computer claimed 50 credits and was granted 20. So what are you talking about?

Check Again Matt

Maybe it changed between when I originally posted and when you saw it, genius.

So , before you bellyache , you should recheck.

How about you stop being a moron?

When I posted it had 0 credit. It was granted credit after that.

This seems to be too complex of an idea for you.

Go start trouble somewhere else.

Time out, both of you.

For the record I have never met either of you, I am not a prior friend of either of you, and I have no side to take in the arguments you have had in the past.

Matt's original post was a useful one - it has pointed out room for improvement in the settings that are used to generate WU. If I am right about that, those tweaks will save the project team manual work in the future repairing broken results.

Jose's reply was inflammatory. There had already been a posting, by me, that mentioned the fact that this project often applies manual adjustments - some of these are in fact done by scripts that run from time to time, others are literally hands on by the project team. An obvious reason for the discrepancy between the value seen by Matt before his first post and the value seen by Jose before his first post.

Jose: you really should have thoght of that possibility, especially after reading my posting.

But also:

Matt: your reply was also inflammatory. You could have put past disagreements aside, and explained politely to Jose that the value had changed since your posting.

Matt gave some good advice in a previous thread about ignoring posts from certain people. I intend to follow it concerning both of you and only the pair of you.

I suggest that Jose follows Matt's advice concerning Matt (as he would have been better to do already).

I suggest that Matt follows his own advice concerning Jose (as he would have been better to have done here).

I will be following my own advice - right after this posting ;-) Both of you are now on my ignore list for a while. I used a T word before, in that other thread. It applies equally to both of you, IMHO. Jose for persistently picking fights and Matt for persistently escalating the invitation.

Finally Jose am I a mod? Yes I do have a mod a/c. That has nver been a secret.

My views here are as an individual participant -- that is why I am posting as R~~ not on my mod a/c.

As a matter of policy I do not act as a mod on an issue where I have already become emotionally involved as I have here, by Jose & Matt's shared kindergaterten pettiness. When I feel I have got myself sucked in, as here, I leave it to other mods to sort out. Including deleting this post if felt necessary.

When posting as an individual participant my postings are equally subject to moderation as anyone else's. And that is as it should be.

River~~

I'll look into possibly modifying the validator to fix this.

hi David,

Just to be clear: it is not the validator - the validator never got a look at the third result here.

All that needs to be modified (if I understand correctly) are the values for the max number of results in the WU type. It is possibly too late for the current work but could be tried as soon as you start the next type of wu.

R~~

Finally Jose am I a mod? Yes I do have a mod a/c. That has nver been a secret.

My views here are as an individual participant -- that is why I am posting as R~~ not on my mod a/c.

As a matter of policy I do not act as a mod on an issue where I have already become emotionally involved as I have here, by Jose & Matt's shared kindergaterten pettiness. When I feel I have got myself sucked in, as here, I leave it to other mods to sort out. Including deleting this post if felt necessary.

When posting as an individual participant my postings are equally subject to moderation as anyone else's. And that is as it should be.

River~~

Woozy Sparky!!!!!!

What caused this reaction? I did not mention you nor your mod status (Although now I am asking myself which one of the mods you are :)] So there is no reason for this angry reaction.

The quoted reaction is flaming too. Worst it was not a reaction to provocation.


Before you point fingers , look yourself in the mirror. BTW all I pointed to Matt before he attacked me is that he got his points. He insulted me, I will not queep quiet when he does.

---

This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato