The hard part is assessing the ramifications of this change.

It's not like we never knew that it would be superior to sort by exact dates and tournament IDs: the year-only-sort was an intentional decision for efficiency reasons. It's *much* faster to index on a two-byte field than to sort on multiple larger fields.

It could be that we might have to wait to upgrade the server with more RAM before we can do this without paying a huge cost in server response time. Or maybe not: we already have a lot of RAM and the index tables currently fit comfortably as things stand.

You can now search the Tournament Index with a free form text query. So you don't need to flip through the pages of chess history any more, just type what you want and up it pops.

The results show up to 100 tournaments sorted from what it regards as "most relevant" to the "least relevant".

A few examples to show how powerful and easy it is:

Tournament Index search: "Leonid Stein"

Tournament Index search: "Wijk aan Zee"

Tournament Index search: "Leningrad"

and believe it or not, you can even make a search like this

Tournament Index search: "TheFocus"

Can you please upgrade <PGN4WEB> to its newest version, <2.80>? You are currently using <1.96>.

In the newest version of <PGN4WEB> clicking the b8-sq gives easy access to the FEN string representing the current position on the board.

Thanks

----

PGN4WEB -> http://pgn4web.casaschi.net/home.html

I've attempted to pull together a decent bio for the late Gyula Sax.

To say the least! Your effort is heroic. So heroic in fact that it's triggered problems that have hitherto gone unseen. I can explain very easily what's going on. How to fix it is a harder question.

What you've done is reached a tipping point. Off the top of my head I don't recall the maximum number of bytes that a biography can be, but it's clear that you've hit that limit. And the software, rather than warn you that sometime is amiss, unceremoniously truncates the biography in the most procrustean manner possible.

This is why your footnote numbers go all the way up to 115 and yet at the bottom the last thing on the biography reads <(101) Paragraph info sourced from Sax’s details provided at> and then it stops in mid-sentence. This is also why there are unusual italics at the bottom of the page: there is an italics tag <i> in the biography that had its closing tag snipped off.

Looking at this biography brings up other questions. For example, what if a new sentence was added near the top, say between footnote 10 and 11, and it also requires a new footnote. Would somebody have to manually renumber over 100 footnotes?

I guess it's also worth mentioning that there are some argue that giving links to raw URLs is a dicey practice to begin with, since links so often become defunct over time. Perhaps the first footnote (for example) could be better written <(1) Member Details USCF http://main.uschess.org/msa/MbrDtlM...> so that if the link ever dies, at least future readers will know what the link used to point to. (Of course, had you done that, even more would have been truncated.)

Just thinking aloud, I can name a few things which might be good ideas if your Sax biography becomes a new standard:

• A larger limit on the size of biographies.

• A wikipedia-like method of inserting footnotes in a standardized manner so the author needn't worry about the actual numbers, but just inserts the source information into the text itself and lets the software present it properly.

• Possibly we could use a way to present extremely lengthly footnote lists in some way that the user has to click to reveal them, or scroll through them, or have them appear in a pop-up window. The footnote content is very important but it's very unlikely that a visitor the page has a dire need to examine it.

• Links to a very select sites already appear in a special format, notably rusbase links will appear as [rusbase-1], [rusbase-2], [rusbase-3], etc. Your biography makes me want to give a similar treatment to 365chess.com. (A great site, by the way. We use 365chess.com to double-check our own data all the time.)

We have every plan of doing that very soon; I'm even in correspondence with the kind author Paolo.

The only reason why it's been put off is because of the recent Java craziness. So many users were suddenly flipping over to pgn4web and simultaneously fighting Java problems, so we didn't want to throw another variable in the mix.

The search is biased towards words that appear in the official name of the event but it also checks words in the article, as well as the player names and a few other secret spices.

<To say the least! Your effort is heroic.>

Thanks.

<So heroic in fact that it's triggered problems that have hitherto gone unseen.>

Sorry about that.

<I can explain very easily what's going on. How to fix it is a harder question.What you've done is reached a tipping point. Off the top of my head I don't recall the maximum number of bytes that a biography can be, but it's clear that you've hit that limit. And the software, rather than warn you that sometime is amiss, unceremoniously truncates the biography in the most procrustean manner possible.

This is why your footnote numbers go all the way up to 115 and yet at the bottom the last thing on the biography reads <(101) Paragraph info sourced from Sax’s details provided at> and then it stops in mid-sentence. This is also why there are unusual italics at the bottom of the page: there is an italics tag <i> in the biography that had its closing tag snipped off.>

<Jess> did a byte count and thought it was way below the limit:

<Blank lines don't use any "bytes" and we get 60k bytes as the limit to these intros. Your entire text (from the edit box) is only 34343 characters, which mans the "byte" number is just a little more than that, nowhere near the limit.>

Is it the url or the brackets that blow the byte count out of the water?

<Looking at this biography brings up other questions. For example, what if a new sentence was added near the top, say between footnote 10 and 11, and it also requires a new footnote. Would somebody have to manually renumber over 100 footnotes?>

No. I would add a 10A insertion. Not the perfect solution, but quite satisfactory in my opinion. It's not likely to happen in this case as there won't be any new events or results added to Sax's bio, for obvious reasons, unless I've left something out.

<I guess it's also worth mentioning that there are some argue that giving links to raw URLs is a dicey practice to begin with, since links so often become defunct over time. Perhaps the first footnote (for example) could be better written <(1) Member Details USCF http://main.uschess.org/msa/MbrDtlM>... so that if the link ever dies, at least future readers will know what the link used to point to. (Of course, had you done that, even more would have been truncated.)>

Good idea, I'll give it the makeover. But I'm not sure I want to do that with 365chess or TWIC. That would double the size of the footnotes.

<Just thinking aloud, I can name a few things which might be good ideas if your Sax biography becomes a new standard:

• A larger limit on the size of biographies.>

Yes!

<• A wikipedia-like method of inserting footnotes in a standardized manner so the author needn't worry about the actual numbers, but just inserts the source information into the text itself and lets the software present it properly.>

If you could do that, wonderful! Saves on artificially numbered insertions.

<• Possibly we could use a way to present extremely lengthly footnote lists in some way that the user has to click to reveal them, or scroll through them, or have them appear in a pop-up window. The footnote content is very important but it's very unlikely that a visitor the page has a dire need to examine it.>

Sounds good.

<• Links to a very select sites already appear in a special format, notably rusbase links will appear as [rusbase-1], [rusbase-2], [rusbase-3], etc. Your biography makes me want to give a similar treatment to 365chess.com. (A great site, by the way. We use 365chess.com to double-check our own data all the time.)>

TWIC too. It's one of the very best resources for researchers on the web.

<In summary: I am very impressed at your output; this is genuine academia and we would be proud to publish it. However, at the same time, you've broken our software that never expected to get input like what you've given it. Not your fault, but I don't have a quick-fix for this one. As it is such an important issue it deserves some serious consideration.>

Like I said, I don't have a quick fix. In the meantime, yes I suppose if you had an account called "Sax Sources" we could have a quick way to link to it, albeit unorthodox. I almost hate to say it, but you could temporarily find a way to shorten the biography until we have the resources to do it properly.

About the byte count: there are really two text fields, the "raw bio" which is what the editors see in the form, and the processed HTML-version. The HTML version can be significantly longer than the raw version, especially when you have links, graphics, etc. I think your citations at the bottom is what's bloating the size of the HTML-version. So while your "raw bio" fits in the 64K limit, the HTML version is being truncated.

I just looked up our database definitions. The field is called a "blob" type (binary large object), and MySQL has four sizes of blobs:

TINYBLOB - 255 bytes
BLOB - 65535 bytes
MEDIUMBLOB - 16,777,215 bytes (2^24 - 1)
LONGBLOB - 4G bytes (2^32 – 1)

Some examples of how to use the Wikipedia system can be seen at:

https://en.wikipedia.org/wiki/Wikip...

<I suppose if you had an account called "Sax Sources" we could have a quick way to link to it, albeit unorthodox. I almost hate to say it, but you could temporarily find a way to shorten the biography until we have the resources to do it properly.>

It's very difficult to shorten the bio to the extent of fitting in all the footnotes. Problem is compounded by the fact that that amount of foreshortening will also mean removing some or many footnotes depending upon which text I chop, and then renumbering the rest, while facing the prospect of reversing that process once the problem is fixed...A simple replacement of saved text instead of reversing the text might overwrite any changes that have been made in the meantime.

I already took out a paragraph from the eulogy, one involving an analysis by Judit Polgár of a game by Sax, and found that it only made a difference of a couple of footnotes.

What I propose to do is retain as many footnotes as I can within the bio and transplant the truncated portion into another account, as discussed. Once you've installed a fix, I could transfer the footnotes back into the bio and you could delete the account.

Would that work?

I know some people don't like long bios as they feel it takes too long for them to scroll down to the kibitzing.

If you allow for long bios, a possible work around might be to truncate the presentation to a few lines, with the option of opening the full bio with a "more" tag.