Primenet web design

[petrw1]

Team name	More Power!!!
Team web site URL	http://www.mersenneforum.org/showthread.php?t=654
Public stats display?	yes
Team created	2008-11-08
Team info last updated	2010-05-05
Team last CPU activity	2011-05-05
Number of members	27
*** Most Recent Activity	2011-05-05
*** Least Recent Activity	2015-07-17

[Madpoo]

Originally Posted by petrw1

Team name	More Power!!!
Team web site URL	http://www.mersenneforum.org/showthread.php?t=654
Public stats display?	yes
Team created	2008-11-08
Team info last updated	2010-05-05
Team last CPU activity	2011-05-05
Number of members	27
*** Most Recent Activity	2011-05-05
*** Least Recent Activity	2015-07-17

I'm not sure where on the site that shows up... At any rate, kind of weird to have a "least recent activity" at all, even if it were correct. I wonder what that would indicate... the first result that was checked in maybe (separate from the "team created" date?

[Madpoo]

Originally Posted by retina

In this example there has been one/two LL test(s) reported. In the two sections that show this/these LL result(s) the user names are different, namely "Carlo Monari" and "ANONYMOUS" appear to have reported the same residue on the same exponent but managed to insert their results into different sections of the DB. User "ANONYMOUS" has a date attached, but user "Carlo Monari" reported the result without any date.

Perhaps these two users live in a superposition of both entities combined. A gestalt of some unknown characteristic. Reporting a result collapses this gestalt into either one, or the other, but never both at the same time.

That's a result of the Primenet v4 logs that were recently integrated.

The "LL" section shows data as recorded in it's own table, which will generally include the v4 ("legacy") user info, and didn't have a date. I guess Primenet v5 is when dates were added to that info.

The "History" section displays data from a different table that has the raw messages sent from the client. George recently tracked down all the pre-2008 logs for those and James parsed them out into a format we could insert into the DB.

Before that was done a month or two back, you wouldn't have seen that info in the history at all...it would have been a mystery where those LL results (or factors, or whatever) came from.

There would be some work to try and get those old log messages to get parsed to show their v4 user name. The raw message contains their login name in most cases, but that's not always the same as their "display name". And if they've mapped a v4 to v5 user we might want to show that... So anyway, for now most of those are just going to show "Anonymous"

George has also mentioned using the time stamps in the LL results of the old logs to update the LL results table. It's straightforward for most of them, I think. There are always going to be the weird ones, like where some user kept submitting the same result (same shift count) over and over. I guess in those cases we'd just take the first one and discard the rest.

That's why sometimes you'll see multiple entries in the "History" section but only one LL result... because the shift counts were the same, and fortunately the web site is smart enough to only show one in a case like that.
e.g. M41803987

Anyway, it's on a to-do list somewhere... match up dates and possibly get user names showing up.

[retina]

Originally Posted by Madpoo

That's why sometimes you'll see multiple entries in the "History" section but only one LL result... because the shift counts were the same, and fortunately the web site is smart enough to only show one in a case like that.
e.g. M41803987

I understand the reason now.

But your example exponent above shows three different dates for the two reported results. I feel a bit like Alice in Wonderland now.

[Madpoo]

Originally Posted by retina

I understand the reason now.

But your example exponent above shows three different dates for the two reported results. I feel a bit like Alice in Wonderland now.

Magic!

Truth is, the dates on the old v4 logs are just based on a (usually) one week period in which all of the logs for that week were aggregated.

In the rare cases (generally in early 2008 when Primenet v5 was being put into play) there will actually be a time stamp on the LL results. And as you noted it may differ from the loosey goosey log dates.

[Madpoo]

I just made a minor tweak to the exponent report page, adding some helpful (?) links.

For example:
http://www.mersenne.org/M36225323

The "LL" text is now a link to the LL specific report page where it has a little more info (any errors reported by the user, and the machine name info).

And if it's currently assigned to someone, the date on that assignment is clickable and will take you to the assignment page for that exponent.

I deliberately didn't make those links obvious... they're the same color as before, but they will underline when hovered. This is the same as the other links on that report like the exponent itself going to the mersenne.ca page, or if it has factors, clicking on the factor will go to mersenne.ca as well.

If people think those links need to be super obvious and jump out at you, I tried a few things like making them always underlined or changing the color, but for now I felt like having it just blend in.

One other minor change was that the hover-text on the dates of different things will drop the milliseconds. It seemed like a useless bit of extra accuracy. If you scrape that page for anything and you simply MUST have it accurate to a few milliseconds, let me know. Otherwise it just added to the page weight a small amount and I trimmed that fat.

[James Heinrich]

Originally Posted by Madpoo

It seemed like a useless bit of extra accuracy.

Can you please take that surplus accuracy and add it to the daily XML dumps, as in the to-the-second accuracy on DateReceived datestamps I requested previously, thanks!

[Madpoo]

Originally Posted by James Heinrich

Can you please take that surplus accuracy and add it to the daily XML dumps, as in the to-the-second accuracy on DateReceived datestamps I requested previously, thanks!

Doh! Oh yeah, sorry, I forgot about that, and after I'd promised to do it quickly. Whoops.

It'll be there for tonight's data dump.

[James Heinrich]

Originally Posted by Madpoo

It'll be there for tonight's data dump.

Thanks

[retina]

Originally Posted by Madpoo

I deliberately didn't make those links obvious... they're the same color as before, but they will underline when hovered.

That only works for people that let webpages set things like that. For people like me the links are just like all other links and styled exactly how I want them.

[Madpoo]

Originally Posted by retina

That only works for people that let webpages set things like that. For people like me the links are just like all other links and styled exactly how I want them.

Wait a minute... you're telling me your browser will ignore the "color: black;" styling on any anchor tags? That's just... I don't know what that is.