Is "reading" absolute? As in:
- "You can learn reading with tsumego, playing 9x9, stuydying joseki or whatever, and apply your knowledge to any reading".

Or there are many types of reading and each situation has its own unique patterns and behaviours, as in:
- "You need to learn reading with tsumego, 9x9, stuydying joseki and whatever."

The basics of Go reading are universally applicable for tree-representable, two-player games. The basics are described independently of Go in theoretical informatics or specifically for Go in the books Tesuji (Davies) / Introduction or - in greater methodical detail - First Fundamentals / chapter 10 Reading.

Nevertheless, specific applications of reading can be performed as different types and there can even be different types for one specific application (such as life + death). However, so far the different types of reading have, AFAIK, not been carefully described yet (except when speaking of algorithms).

E.g., a type of life + death reading can make assumptions such as reading only within a locale or can use short-cut tools such as nakade knowledge. One can set also side conditions, e.g., first verify that a group is alive at all, then refine reading to find life while maximising local territory.

Typically, basic reading during the opening is quite different: one reads only a few moves deep and prunes all deeper iteration. (Extended, possibly deeper reading can clarify additional questions.)

E.g., Monte Carlo programs use a very different type of reading: they create samples of variations. Human players can also use this method but with only a few samples, which are generated differently.

Reading can be accompanied by aims, filters, knowledge relations and other means. E.g., status reading can filter all variations, in which the moving player fills an intersection of a two-eye-formation. In practice, this filter is good; in theory, it can lead to overlooking perfect play (under rules with superko).

Not only are there types of reading but the variety of types and methods is rich. Programmers speak of different (types of) algorithms.

When studying joseki, it helps if you can read the language the comments are written in.

_________________
http://tchan001.wordpress.com
A blog on Asian go books, go sightings, and interesting tidbits
Go is such a beautiful game.

As I have mentioned in e-volume 3 of Shuei's games, there was a brilliant series by Murashima Yoshinori around 1951-52 when even top pros had to write their own articles. It was called "How to nurture reading strength".

There was a brief but superb and insightful description of the fundamentals, which involves how to think first about shapes in open-field fights. This knowledge is then transferred to the corner, i.e. josekis, and compared and contrasted. In modern terms he shows how to prune the tree, order candidate moves, and so "avoid futility" of analysis.

Evaluation of lines was important, of course. Three special insights were: (1) think about solid and high rather than thickness and profit (not just in joseki), (2) when comparing open-field shapes to corner shapes it is the second line in josekis that needs special attention (because of the edges), and - best of all - (3) you assess and understand josekis not by counting profit and thickness but by understanding and making the right "fighting shapes".

It's nearly all text/theory and I could have done with more examples, but it had a huge impact in helping me understand (at long last!) some of Shuei's joseki play (and Murashima also does a similar job vis a vis Go Seigen).

As I said in the book (coming soon), Murashima was one of those prized experts who have the ability to say something new and profound and yet leave the listener believing it was so obvious that he knew it all along (only somehow he never quite got round to telling everybody). There should be a word for experts like that. Does anyone know one? Or want to invent one?

Guru.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

There I disagree. Josekis can be evaluated very well by counting numbers of played stones, territory and influence stones (provided one uses such a neat theory as mine). Assessing josekis by understanding shapes (not only fighting shapes) and creating josekis by making meaningful shape moves is another good method, which I also advocate (see, e.g., Joseki 1 chapters 9+10, Joseki 2 strategic lines chapter, Joseki 3 chapter 2).

Why would one dismiss one of both methods? Using both creates more confidence!

Did Murashima attempt to assess thickness by numbers and how?

If Murashima did not assess thickness by counting at all, then (3) is not his best insight but, quite contrarily, is a bad excuse for not knowing how to do it. In particular, quickly assessing thickness by counting can be a very efficient means to prune reading. The question occurs why then one would call his series of articles "brilliant".

If he did assess thickness by counting, then the quality of his assessment method determines whether his statement has at least some justification.

Therefore let me ask again: Did he assess thickness by counting and, if yes, how?

(Counting territory (profit) is a means of pruning more difficult reading and a basis for strategic judgement, so his statement about not counting territory does not make sense.)