saintbletch

Kilmarnock. No, I don't even know him. Dundee. Did you?

great, big and cock*? *verb form

No 3B, you are not worthy, and yet as a measure of the man I am I will still deign to interact with you and patronise both you and this place by my presence.

Wouldn't that mean that you were writing those things about other people - in reply to me, perhaps?

So back in your day what did you used to call a dictionary?

Glad you're enjoying the hilarious fun that can be had with words, Halo. Toke is too, but he's a little reticent about admitting it. Well, now you ask Halo, you can do all sorts of analysis on the text. You can form impressions and trends from the text, rather than being able to draw cast iron conclusions. So alas, nothing I do will end up getting some of the main-boarders detained for 28 days under the influence of messrs valium and zanax. What I can do is gain an idea of the sentiment and subjectivity of a post, poster, thread, forum, etc. This means I can then compare each of these results with other results. So questions such as "Is Alpine_saint more negative than the notional average poster", can be 'answered'. I can also then combine things and compute the 'average' sentiment in specific noun-phrases - such as whether the forum is on 'average' positive when I detect the presence of specific tokens - such as "cortese", "lambert", "lallana" and so on. I can then look at that temporally - so I can see if the attitude towards "Hooiveld" has changed over time, and at what time it started to change. I can also look at things like verb usage and correlate that with Saints' results, to see what sort of language gets used after wins, losses, draws, etc. I can ask questions such as which posters have contributed the most negatively to threads that got locked. But all of this comes with a mass of caveats. For example, not every poster uses the Queen's which leads to less accuracy, and so many memes, some specific to TSW, need to be classified (woot, nuns, facepalm, clappy, clappie, clappies, happies, bedwetters, etc.). Also, this might come as a surprise, but some posters dominate discussions. Having a significant portion of a thread contributed by one poster can really skew results. Another issue is the use of sarcasm. I am also blind to the real meaning of posts by Turkish and CB Fry because they unremittingly rely on sarcasm (woosh!), which would skew my 'understanding' of their 'work'. It's possible to get around all of these problems, but it involves a lot of effort in training a program. This means looking, sentence by sentence, at posts and manually classifying them. Once done, this manual classification can then be taken and used in an automated way. This is all a branch of computer science called Natural Language Processing. It, and a subset called Sentiment Analysis, are used to mine the words we excrete into electronic social interactions every day. Organisations and companies want to know what we think about their products, services, competitors, government policies, opposition, secret services, etc. Work in this area is really taking off, and I'm benefitting by standing on the shoulders of some really smart people that have made this accessible to hobbyist such as me. I'm not a proper programmer any more, I used to develop for a living many years ago, so this is all a bit of a sideline now.

OK, just this once and mainly to ****-off Toke by seeming to favour you over him, I'm going to allow that suggestion even though you didn't quote the post in question. You're wrong anyway, but, and look at me being a tart here, post 299 isn't BTT... (The power of the ellipsis, eh?)

Tokyo is fertile and Bear is fecund Neither could get it up So I want a refund.

Not so, notnowcato.

Dear Bear, at 1432 this afternoon, this post simply asked if post #392 is rallyboy. If you plan to edit it every time somebody has a guess in order to make it appear that you guessed first, I will slap your legs. Rallyboy was correct though, so I will give you the 6 points back that you lost earlier in the day.

Not correct.

I see what you mean, Bear.

Nope.

Well I don't know if this poster has come back from the dead or not, but it is not Glasgow's stalker.

No.

No.

No, not Hatch.

Nope.

Oh, we're at the clues stage, 3B. When you start a book, do you skip to the final page first to see if you'll enjoy it? Pah! Oh, well. One is a TMS regular, one is banned and one, shame on me, was a completely new name to me (that's why it is fiendishly difficult/impossible).

I'm a frayed knot, pap.

Correct, nowcato not. #300 is not 3B.

Too late papster.

Kerching! With reputations as large as that, it had to be Spudders. notnowcato has one point.

Nope, the following are still not solved. Posts #289, #299, #300 and #308.

Could be. Which word cloud are you referring to, Toke? Are you sulking because pap can follow rules and you can't?