Primenet web design

[cuBerBruce]

Originally Posted by Madpoo

If you sort just by days to go it would probably do what you're after (i.e. it would be sorted by that, and the estimated completion column would also be sorted correctly.

NOPE. If you looked at the graphics I provided, sorting by days to go clearly doesn't get it right either. Back when both date and time were included, the exponents would be in the correct order when sorting by estimated completion.

If I do a 2-tier sort, I at least get the exponents sorted correctly as far as one can tell from the information displayed. I believe it was someone else (i.e., not Madpoo) that eliminated the time of day info from being displayed int the Estimated Completion and other date fields.

[Madpoo]

Originally Posted by cuBerBruce

NOPE. If you looked at the graphics I provided, sorting by days to go clearly doesn't get it right either. Back when both date and time were included, the exponents would be in the correct order when sorting by estimated completion.

If I do a 2-tier sort, I at least get the exponents sorted correctly as far as one can tell from the information displayed. I believe it was someone else (i.e., not Madpoo) that eliminated the time of day info from being displayed int the Estimated Completion and other date fields.

I remember removing the "time" part from some report or another, maybe not that one. Whichever one I did, the full date/time is still there when hovered over, and I think I setup the table sorter to look at that hover info for sorting.

I could look back at whatever I did for whatever report that was... once I remember which one it was. :) It was SOOO long ago (not).

[cuBerBruce]

Originally Posted by Madpoo

I remember removing the "time" part from some report or another, maybe not that one. Whichever one I did, the full date/time is still there when hovered over, and I think I setup the table sorter to look at that hover info for sorting.

I could look back at whatever I did for whatever report that was... once I remember which one it was. :) It was SOOO long ago (not).

This is the Account Assignment Details page (from the menus: Account/Team Info -> My Account -> Assignments) that I'm talking about. Hovering over the dates does not seem to cause any other info to be displayed. (The row where the mouse cursor is gets highlighted, but that doesn't require hovering.)

It's my recollection that the removal of the time information was done before the change to the new GUI and the new server. Someone felt that four date fields all including time information made the table unnecessarily wide. Personally, I would like an option of at least being able to get the Estimated Completion column to include the time (and not simply getting it by hovering), especially for assignment within a few days of completing. I don't know, maybe I'm the only one who misses this functionality that was removed?

[Madpoo]

Originally Posted by Madpoo

I remember removing the "time" part from some report or another, maybe not that one. Whichever one I did, the full date/time is still there when hovered over, and I think I setup the table sorter to look at that hover info for sorting.

I could look back at whatever I did for whatever report that was... once I remember which one it was. :) It was SOOO long ago (not).

Okay... I think I was a little confused earlier... I was thinking of the generic "assignments" page where you can lookup assignments in a range, but you were talking about the individual assignment page, /workload/

I updated that to include the date as well as hour:minute in an attribute "data-text" that the table sorter will use if present. It keeps the display minimal but lets it sort properly. I guess if you really wanted to, you could look at the page source and get the hour:minute for the actual dates. I guess some people might manage their assignments to the point where hour/minute granularity from the report page is nice to have.

I don't remember if it was this report or another, but I experimented with having date and then a linebreak in the cell, then the time. But someone said it messed up the way they hit that page and parse the content... that might have been some other report though, not this one.

I'll have to look at rolling this same change to the team workload page so it works similarly, but no time for that tonight... it'll have to wait.

EDIT: Oh, I just looked a the team workload and it's showing the full date/time for assignments. I kind of remember now that to get it to look good, I had to use smaller fonts otherwise cells will start wrapping funny. So I guess if you wanted to see the times, if you're not already in a team just create one for just you and look at the /tworkload/ page. :) I can try looking at squeezing more info into the account assignment page, but I won't promise anything. :)

[Primeinator]

Originally Posted by cuBerBruce

Someone felt that four date fields all including time information made the table unnecessarily wide. Personally, I would like an option of at least being able to get the Estimated Completion column to include the time (and not simply getting it by hovering), especially for assignment within a few days of completing. I don't know, maybe I'm the only one who misses this functionality that was removed?

I would also like to see this feature brought back (if feasible).

[Jayder]

Thanks for your reply, madpoo. I like your ideas. Basically, I'd like to see which numbers have had the least work done on them, as you say. I'm going to change my suggestion. My wish would be to be able to see the exponents with the least TF work done on them, the least P-1 work done on them, the least ECM work done on them, and the combined factoring work. By no means does it have to be anything close to the below, but it is what I am currently imagining. I'm sure somebody could come up with something that makes far more sense.

Exponent       TF work        P-1 work       ECM work       Total work
----------------------------------------------------------------------
4,500,007      1.6213 GHzd    0.1253 GHzd    1.1087 GHzd    2.8553 GHzd

I'm not sure that it makes sense to have ECM work in GHzd or if it would be better in whatever internal measurement you use/T-levels.

If it's decided that this is worth implementing, I envision being able to search within a range of exponents, and not just the whole database. Another possible column could be the probability of having found a factor, if feasible.

This may be better suited for something like James' site, but I've bothered him enough.

Disclaimer: I am open to the idea that this is a bad suggestion. I also recognise that continuing work on cleared exponents is distasteful to some. Please be kind. (I feel that adding this line is necessary. I know that many here love to belittle others, and I am not fully confident in the value of this suggestion. This disclaimer is in no way directed at madpoo. Isn't it sad that I feel the need to say this?)

[James Heinrich]

Originally Posted by Jayder

This may be better suited for something like James' site

Since a long time ago I've had the Poorly P-1 factored page which used to be very useful in finding exponents with badly-done P-1. Literally thousands of (small, below 10M) exponents had P-1 redone and hundreds of new factors were found.
It's still somewhat useful if you're looking to redo some P-1 work that was badly done (B1=5000 or something silly). It won't give you any info about ECM or TF though (I also have a page for less TF than expected but a lot of what's shown there is borderline cases where the rough formula I use to calculate target TF level doesn't match up exactly with the precise formula that GIMPS uses.)

[Madpoo]

Originally Posted by Jayder

Disclaimer: I am open to the idea that this is a bad suggestion. I also recognise that continuing work on cleared exponents is distasteful to some. Please be kind. (I feel that adding this line is necessary. I know that many here love to belittle others, and I am not fully confident in the value of this suggestion. This disclaimer is in no way directed at madpoo. Isn't it sad that I feel the need to say this?)

I think it would be a nice idea to have some kind of "poorly factored" report, along the lines of what James also replied.

The question I'd have is, how to determine poorly factored P-1/ECM/TF work. I know it wouldn't just be a matter of looking at the total ECM work done so far, because I'm sure there are plenty of larger exponents that have had a smattering of ECM work done (a curve here and there at some small bounds), and yeah, I guess technically they're "poorly factored" from an ECM point of view, but it would probably be better to compare the ECM effort to the exponent size, for instance.

For TF work, James might be able to help me figure that out based on what's available in the GIMPS data, since he has something like that on mersenne.ca. By the way James, I've been meaning to email you and see what it would take to refresh your data... I ran across some exponents recently where it seemed like maybe the info on mersenne.ca wasn't quite up to date. But I digress...

Maybe it could be some added option on the manual assignment page, to get "poorly factored" ecm/p-1/tf work... I'm just spitballing here, but something like that? A report of the same poorly factored stuff for analysis is good, but if you want to get assignments for that type of thing, seems like you'd want that option on the manual assignments page just to keep track of it all and avoid people doing overlapping work?

I'd defer to George on that idea. I don't think he minds if I'm updating reports and layout stuff but when it comes to assignments I definitely don't want to mess that up.

[James Heinrich]

Originally Posted by Madpoo

For TF work, James might be able to help me figure that out based on what's available in the GIMPS data, since he has something like that on mersenne.ca. By the way James, I've been meaning to email you and see what it would take to refresh your data... I ran across some exponents recently where it seemed like maybe the info on mersenne.ca wasn't quite up to date. But I digress...

GPU72 has long taken the lead on knowing what exponents best need TF, at least as far as optimal GIMPS throughput.

As for my data, it's never quite current, but I missed at least 3 weeks of "recent results" when you changed the report page and I failed to realize I was parsing the HTML and not text version, and my spider failed to tell me it was failing...
If you want to finish 24-hour historic XML reports that would be useful to me for ensuring I'm not missing any data.

[Madpoo]

Originally Posted by James Heinrich

GPU72 has long taken the lead on knowing what exponents best need TF, at least as far as optimal GIMPS throughput.

As for my data, it's never quite current, but I missed at least 3 weeks of "recent results" when you changed the report page and I failed to realize I was parsing the HTML and not text version, and my spider failed to tell me it was failing...
If you want to finish 24-hour historic XML reports that would be useful to me for ensuring I'm not missing any data.

Yeah, I think the daily XML thing is probably going to be "easy enough". I have the TSQL written, so it's more a matter of double-checking the output and then figuring out how best to script it to generate and dump an XML for each day... probably going back 30 days to start and check performance, then schedule it to go back however far for the rest.

I don't know if MySQL can take something like a BCP dump from MSSQL... in which case we could probably arrange a transfer that way... factor info for starters (exponent, current TF depth, P-1 bounds, all known factors, etc).

It'd be a big set of data and BCP is probably the optimal way of dumping it. I'd refer to it as "the firehose" since it's a large set. Not just James, but as long as any sensitive data is masked/removed (full residues, private user info, etc) I don't think George would mind if others had access to it.

It'd be "point in time" because it's not something that could feasibly be kept up to date on any kind of regular schedule, but for an initial "import" or to setup a temp table and verify your existing data, it seems ideal. I'd discuss it offline, but like I said, other people who crawl the various stats and aggregate/analyze the data in fun and different ways could benefit.

[James Heinrich]

Originally Posted by Madpoo

I don't know if MySQL can take something like a BCP dump from MSSQL...

Supposedly: yes. I've never seen a BCP file so I can't say I've tried it, but it would seem that it can be imported, at least if the BCP file contains data for only one table. Perhaps you could dump 10 rows of some relevant table and email me a sample BCP and I'll let you know how it goes.