[QUOTE=Mark Rose;384680]I'm a canuck and I didn't understand that exchange at all.[/QUOTE]

[QUOTE=James Heinrich;384681]Me either. But Chris is his own kind of ex-.ca-in-a-warm-place, perhaps his humour has softened in the sun.[/QUOTE]

What's warm?

[QUOTE=TheMawn;384746]What's warm?[/QUOTE]You know in July when the ice gets mushy? Imagine that all year long.

[QUOTE=James Heinrich;384751]You know in July when the ice gets mushy? Imagine that all year long.[/QUOTE]

When it was explained it was a Canuck joke, all I could think was that I missed that episode of Red Green or SCTV. :smile:

[QUOTE=Madpoo;384727]...
The same should apply to pretty much all the other tables. Once I can get the arrows added back to the column headers without messing up the rendering (it renders, then resizes the columns once the theme loads since the arrows add width, which I didn't like) you'll be able to see it better.[/QUOTE]

So... I think the problem I had with tables doing a re-paint when the table theme loaded was probably specific to the really LONG tables (top 500 producers, etc). It doesn't seem to be much of an issue with the CPU benchmark table.

And honestly, now that I look at the CSS being applied, I'm not sure I know why it was repainting... I don't think it's the arrows at all since they're just applied as a background image for the header cell. Oh well.

I'll add that in for the benchmark table, so be looking for that soon. I need to tweak it a bit too so that the default sort order is different for the different types of reports available. I like having the basic page (showing all CPU models with whatever options) sort initially by fastest -> slowest, but that 2nd column is different for the tables like "show only my benchmarks" or if looking at just a single CPU model. :)

I'm going to check in with George on how some of the benchmarks are stored since I'm trying to figure out a better way to display the info. Maybe it can be presented in a easier to digest format, I just need to wrap my head around all the raw data.

[QUOTE=James Heinrich;384751]You know in July when the ice gets mushy? Imagine that all year long.[/QUOTE]

Wow. MIND BLOWN.

Factoring Limits Report broken
 

The factoring limits report is broken (sort of). Using the url [URL="http://www.mersenne.org/report_factoring_effort/?exp_lo=332192831&exp_hi=332599999&bits_lo=0&bits_hi=999&B1=Get+Data"]http://www.mersenne.org/report_factoring_effort/?exp_lo=332192831&exp_hi=332599999&bits_lo=0&bits_hi=999&B1=Get+Data[/URL]
there is a problem. The table returned no longer follows normal table construction, i.e. 1 header row with column titles followed by each 'record' in a separate row.

The problem is here:
[CODE]<th>How far factored[COLOR="Red"]<br>[/COLOR](2^n)</th>[/CODE]
That <br> causes the (2^n) to become its own line. Importing this table now fails (with out further coding on this end). I can fix my macros etc. But I think that keeping all tabular reports as standard as possible is a good thing.

What say ye gentleman?

[QUOTE=Madpoo;384754]...
I'm going to check in with George on how some of the benchmarks are stored since I'm trying to figure out a better way to display the info. Maybe it can be presented in a easier to digest format, I just need to wrap my head around all the raw data.[/QUOTE]

Still trying to re-imagine some things... have a few ideas.

For now, enjoy the benchmark table with up/down arrows showing up now for sorting, and I threw in a bonus, being able to filter on the CPU name, or by program, computer name (on your own benchmarks). Just start typing something into the filter box and it'll show matching rows.

[QUOTE=Uncwilly;384760]The factoring limits report is broken (sort of). Using the url [URL="http://www.mersenne.org/report_factoring_effort/?exp_lo=332192831&exp_hi=332599999&bits_lo=0&bits_hi=999&B1=Get+Data"]http://www.mersenne.org/report_factoring_effort/?exp_lo=332192831&exp_hi=332599999&bits_lo=0&bits_hi=999&B1=Get+Data[/URL]
there is a problem. The table returned no longer follows normal table construction, i.e. 1 header row with column titles followed by each 'record' in a separate row.

The problem is here:
[CODE]<th>How far factored[COLOR="Red"]<br>[/COLOR](2^n)</th>[/CODE]
That <br> causes the (2^n) to become its own line. Importing this table now fails (with out further coding on this end). I can fix my macros etc. But I think that keeping all tabular reports as standard as possible is a good thing.

What say ye gentleman?[/QUOTE]

Well... I wouldn't say it was "broken" to put a break in a table header. You can put all kinds of stuff inside a <th> or <td> :)

But I'll grant you, it did break what you're doing with the data. What you might be after is something that would be sure to strip HTML markup from inside the bits you're after. In that particular table it should only be that <br> in the header.

I added that <br> for one reason... the data in that column is a 2 digit number, but the header is really long. It looked goofy. LOL Just this really long info up top, little tiny data. Is that the table equivalent of a mullet?

Some other tables like top producer lists will have anchor tags since some users have a website, and the username is clickable to that link. Or there could even be some spans inside the data to mark up some italic/bold for parts of the text.

I would suggest we look at some better options... I was aware that there were some of y'all who were crawling these reports and parsing data out of it, which is really cool, I think. Unfortunately I wasn't aware of all the different ways people are doing that, and since there's no API or anything for accessing this data, I think the people doing this have just worked their process around whatever the web page looks like, which is NOT the way we should be sharing data, in a perfect world.

My suggestion is thus to create an XML feed out of that same data which would be independent of any website layout changes, table modifications, etc.

Something similar to the thresholds XML I setup for Chris yesterday (still not scheduled, but I did update it again manually today).

If these types of feeds prove popular enough we should probably standardize the schema and make it extensible so I'm not creating totally different XML formats for each report... I've never defined a schema myself except some small little things. Usually I'm given the schema and spit out the data to fit. But hopefully I could come up with something that works reasonably well. So long as I can setup a good hierarchy, it should fall into place, right? :smile:

[QUOTE=Madpoo;384624]Here's where that # comes from (the divisor in that percentage):
...
divided by the magic fudge factor 1.68[/QUOTE]

I had no imagination that you would expose such inner-sanctum arithmetic!
Perhaps that's an inspiration factor to whip up the masses to try even harder to red-line your registered equipment at disposal!
Or is there some conversion happening?

[QUOTE=Madpoo;384764]But I'll grant you, it did break what you're doing with the data. What you might be after is something that would be sure to strip HTML markup from inside the bits you're after. In that particular table it should only be that <br> in the header.[/QUOTE]I am using the excel webquery feature. I have tried several different settings. They all seem to see the <br> as a new row. Excel does get the (2^n) into the proper column. Can you try using a <pre> tag for that cell? I think that excel can be set to ignore the formatting.

[QUOTE=Uncwilly;384760]The table returned no longer follows normal table construction
What say ye gentleman?[/QUOTE]I say if you're spider-parsing that data you should be parsing the plain-text version and not the HTML version.

Plain-text wrapped into the webpage:
[url]http://www.mersenne.org/report_factoring_effort/?exp_lo=332192831&exp_hi=332599999&bits_lo=1&bits_hi=99&txt=1[/url]
Plain-text all by itself is what you should be parsing:
[url]http://www.mersenne.org/report_factoring_effort/?exp_lo=332192831&exp_hi=332599999&bits_lo=1&bits_hi=99&txt=1&download=1[/url]