Website founded by
Milan Velimirović
in 2006

11:12 UTC
ISC 2022



Remember me

Forgot your
Click here!
to create your account if you don't already have one.

Rating lists


MatPlus.Net Forum General Codex Article 23 on correcting & versioning Prior Art
You can only view this page!
(1) Posted by Andrew Buchanan [Saturday, May 15, 2021 15:57]

Codex Article 23 on correcting & versioning Prior Art


Looking at

At the bottom, it says in a text comment field "version of Valerij Smirnov" (2010), which raises various points.

Yes the cooked original was by Niemann, but Smirnov's Meredith correction is very nice and light, with many more synergies than the original (or the clunky but correct one by Bulavka).

Article 23 of our Codex states:

(1) If a published chess composition is found to be unsound, it loses its priority date unless a correction is published within three years after the publication of the unsoundness.

(2) The author of a chess composition which has been published in unsound form retains the following rights:
(a) The right to correct the composition himself, and
(b) The right of being cited as author if a correction is made by someone else.

It is recommended that a correction made by someone else ("B") should, if practicable, be published in agreement with the author ("A"). B’s name may also be mentioned. The following formulae are used (in order of increasing originality of the correction – which is a matter of personal evaluation): "A, correction"; "A, correction B"; "B after A". The correction should be published as a joint composition only if A agrees. This note also applies to improved versions of correct chess compositions.

This still isn't quite clear to me. What does "joint composition" mean? Just "A&B"? Or does "A, correction B" fall into that category? What if A is deceased? What are the options for B then? Can B only unilaterally declare "B after A", because it's not "joint", even though that awards A less recognition than perhaps he deserves. Etc.

Moreover, the whole business of versioning is not scalable. By which I mean that if you have "A version B", and C comes along, you cannot reasonably have A version B version C. In the end attribution has to thankfully collapse into just a bunch of names at the top, and that's what seems to happen in practice, but that collegial act of sharing and community is not even mentioned by Article 23, which to me does not seem balanced in its treatment of the different stakeholders.

As is often the case, the Codex provides a superb first draft, but if someone has a "correction/version" :-) of Art. 23 which fills in some of the holes, I'd appreciate it. And then we can look at how the databases can reflect that best.

And then for all of these options, what can be done in YACPDB? In the "A, correction B" case, B (although subordinate) seems to me to deserve more than a corner of a comment field. (And Smirnov's is a correction, not just a version.) Also, it should say "version/correction by", rather than "version/correction of", which incorrectly reverses the sense.

In PDB it is often clearer. See
(1) All contributors appear in the author field, which is drawn from the multi-language, curated, searchable list of authors, but the usage does *not* imply "main author".
(2) If best practice is followed, the After field (also searchable) contains the creative attribution, which may be one of the specified formats found in the Codex, to clarify the agreed or guessed relationship (here "JN, correction VS").

(Read Only)pid=21149
(2) Posted by Kevin Begley [Sunday, May 16, 2021 01:59]


You have my thanks for exposing a number of important gaps in our codex.

This seems the ideal place to ask for any suggested repairs (anyone?), and to discuss the merits of any methodology presented.
I hope these issues can be resolved promptly, so as to minimize the number of problems which may require alteration.

From my perspective, it's more important to provide a reference to the previous work(s) -- as it seems insufficient to reference only previous contributors.

If the problem community could establish a universal problem database, wherein every published problem is assigned a unique identification, we could more logically declare "Composer C, after problem X" (wherein problem X may be composed by Composer B, after problem Y; and problem Y may be composed by Composer A).

This would seem a more logical way to refer to any prior work upon which the original rests. Furthermore, such a reference would remedy any confusion as to the specific work. That said, I do recognize that such a reference may fail to provide due credit above the diagram of this new interpretation/version.

In music, you will often encounter performers who cover a specific cover of a song. It seems to me, best practice is generally to give credit both to the original artist, and to the cover artist (if the cover provides a significantly distinct interpretation).
But, if we can reference the cover song directly (as a universal problem database might allow), and anyone may locate that work, one could reasonably argue that it is both sufficient and efficient to provide only the database identification for the cover (which will provide a database identification for the original).
(Read Only)pid=21151
(3) Posted by Hauke Reddmann [Sunday, May 16, 2021 10:37]

I think music does not compare well - even the most faithful cover
will be different enough (just look at all the ripoff lawsuits,
"hey, he used ABACAB in his melody too!"). A minimal change
("hey, the bridge before the last 'It took a lost weekend' should
not end abruptly but fade out a little, Lloyd!") simply does not
occur. Thus the whole can of worms of what constitutes an improvement
is specific to our business. My alternate suggestion thus would be
patent law, where small changes and improvements may be relevant.
(Read Only)pid=21154
(4) Posted by Andrew Buchanan [Sunday, May 16, 2021 14:16]

Thanks for the helpful responses. I've requested through Brian Stephenson that the Codex Rules Committee consider reviewing Art. 23. I've made my points in the base post - nothing to add for now.

We spend so much time optimizing our compositions, adding every last nuance, that it does make sense to pay a little care to some of the ground-rules which allow our activities generally to proceed so smoothly. Definitely, the Codex should evolve only by incremental changes, rather than big-bang v2.0 etc. It's largely really good. And it's magnificently *short*.
(Read Only)pid=21155
(5) Posted by Viktoras Paliulionis [Sunday, May 16, 2021 15:24]

I’m not sure if any painter who has reduced unnecessary details or corrected inaccuracies (in his opinion) in a painting by Rembrandt can call himself a co-author.

Different captions mean different relationships with authorship. A&B usually means that both co-authors have equal rights. There is more clarity in the Russian rules for chess composition ( I have translated some of their points using Google Translate:

- The publication of original collective compositions should, if possible, be agreed by the co-authors.

- In the case of purposeful processing by author B of the composition of author A, author B himself determines the degree of improvement. With a slight improvement, the new composition is published under the name "A", below which indicates: "Version B" (Переработка = Reworking). Such revisions are not eligible to participate in competitions.

- If the first composition is significantly improved or changed, the new composition is published under the name "B", below which it is indicated: "After A". These revisions can take part in competitions, and the author is obliged to indicate the source of borrowing. By agreement between the authors, such a composition can also be published as a collective one.

- An incorrect composition can be corrected by the author and published in any printed (or electronic) edition (preferably in the same one where it was first published) or repeatedly (only once) to take part in the competition as an original work. If the author has not corrected the composition within three years, then another composer, after this period, may publish his revision in accordance .... New composition (correction or revision of the first one) can be published as a collective work of the authors with their mutual consent.
(Read Only)pid=21156
(6) Posted by Kevin Begley [Sunday, May 16, 2021 17:19]


To be honest, I'm not sure it's possible for the Codex to legislate honorable behavior.

When computers first appeared on the solving scene, many prize winning compositions were found wanting some corrections.
For the most part, problemists (quite honorably) asked little (or no) credit for corrections offered, save when corrections proved exceedingly elusive to others.
The ingenuity necessary to make some corrections can prove more demanding than developing the original scheme -- some do deserve credit for repairing a Rembrandt (and a few may deserve more credit than Rembrandt himself).

If the 8-man EGTB were released today, any study published prior to 2018 (especially prize winning studies) might be considered open season.
In the absence of definitive legal rules, problemists are perhaps more likely to take the most honorable path.

Today, computers generally only overturn a subset of problems -- such as: very long problems (selfmates, proofgames, series-movers, etc), and
problems employing a new fairy element (or a novel combination of fairy elements).
Quite naturally, this subset of problems tends to be the most interesting (and not only for ambitious solvers).

There are two dangers for the Codex:
1) it may provide too much protection for the original composer (that is: it may provide authors a mechanism to secure an extended period of effective monopoly over some specific thematic content -- either by producing a cooked version which triggers a repair period, or by producing a version capable only of withstanding an immediate cook), and
2) it may provide too little protecting for the original composer (that is: it may provide a mechanism for others to pilfer creative content by suddenly revealing cooks, and their own originals, immediately following the established correction period).

You have no idea how much I hate to admit this, but it may well be possible that an absence of clarity on these matters might actually benefit the community (a lack of clarity might best encourage problemists to behave most honorably, if an intentional vagueness compels each individual to more carefully weigh their own contributions).
I especially don't like saying this because we all know such a house of cards is destined to collapse entirely, the moment it is confronted by an unscrupulous wind.

Until that foul wind appears, problemists tend to behave as though it is always in their best interest to avoid any intelligent planning.
I'd wager heavily on the status quo being renewed.


Great points about why we don't want to model our Codex after modern music (you sold me on the downbeat of Ice-Ice-Baby).
Unfortunately, Patent Law has already spoken (in the case of Bobby Fischer vs the World), and it established no means to claim any chess position.
(Read Only)pid=21157
(7) Posted by Viktoras Paliulionis [Monday, May 17, 2021 00:27]


I believe that the codeх should clearly define four types of authorship (for the second composer):
A & B
B (after A)
A (version by B)
A (correction by B)

These types of headings were used in the past, even before all codexes, and were understood intuitively, but today not all people know their differences.

If you have significantly improved Rembrandt, you can write "B (after Rembrandt)". In this case, your problem will be considered original and will be eligible for the competition. But you can't write "Rembrandt & B". Likewise in case of fixing the problem.

I agree with you that there is currently no need to specify a correction period in the Codeх.
(Read Only)pid=21158
(8) Posted by Siegfried Hornecker [Monday, May 17, 2021 02:53]

My "Study of the Month" article that I plan for this month also talks about priority (among other things). I wonder if I should postpone it for a month, write something else quickly, and then include the ideas from this thread in my article.

Here is my text, as an exclusive preview (could be subject to change at any time), as I hope it might bring an interesting aspect to this discussion:
What is described as "priority" means that only the first published endgame study that shows a specific position and continuation can claim originality. Even the most beautiful endgame study will be disqualified from a tourney if it was published before, possibly by another author. Cases exist where a game anticipated an endgame study idea, but also it is legitimate to publish an endgame study with an idea from a game if that idea was not played, or if the about to be published study adds further innovation.

EDIT: I concluded I probably lack the ideas for another article, so I finish this one instead.
(Read Only)pid=21159
(9) Posted by Hauke Reddmann [Monday, May 17, 2021 11:34]

@Siegfried: You remind me of a certain joint Moscow 1. Prize... :-)

@Kevin: Whereas I would not suggest to outright hire a patent lawyer
to sort out priority claims :-), I just modestly suggest that patent law
might be *inspirational* for the debate. After all, your a) and b) is
exactly what keeps patent law makers rotating: Protect the mental
property of inventors! Don't stifle progress! And somehow make the
both not clash!
(Read Only)pid=21160
(10) Posted by Michel Caillaud [Monday, May 17, 2021 18:18]

For the main topic, two "historical" examples from FIDE Album 2004-06 to think about :
-A150 : I remember that Milan was happy that this problem provided the missing half point granting Matti Myllyniemi (died 1987) the posthumous GrandMaster title.
I like to think that Milan followed the guidelines of his heart rather than the guidelines of some Codex.
-H9 : this was submitted to the Album as "Aleksej Troitszky, corrected version by Dmitrij Baibikov", with some quotations; "the first retroproblem with phases transformation" (Nikita Plaksin & Andrey Kornilov, 1997); "this problem, which Troitzky himself consider his "lovely" retro, should attribute to number of several unsurpassed masterpieces" (Moisej Neiman, 1926)
I don't like to think of what could have happened between the work of the judges and the Album publication.
(Read Only)pid=21161
(11) Posted by Joose Norri [Wednesday, May 19, 2021 08:54]

Just a mention, regarding A150 in the 04-06 album. Milan wrote a short article explaining how he and Myllyniemi had composed it together 20 years earlier, being a version of an earlier problem by MM which had a thematic impurity. Milan forgot about the problem and found it in his papers...

So this is not very relevant in this connection.
(Read Only)pid=21163
(12) Posted by Viktoras Paliulionis [Wednesday, May 19, 2021 12:08]

The question is whether the new practice in PDB of writing the author of the version in the heading as a co-author is correct (e.g. P1389613).
What if H9 were submitted as "Alexei Troitszky & Dmitry Baibikov"?
(Read Only)pid=21164
(13) Posted by Siegfried Hornecker [Wednesday, May 19, 2021 12:24]

Out of respect for Milan, I will strongly hold back on the words I would say.
But the idea that you become a grandmaster because your friend publishes a joint composition with you decades after your death is just... weird.

Matti was a grandmaster before, it was just not officially recognized. That is my takeaway: Titles might indicate your strength as a composer or judge, but they are just that: Titles. They have no real value, except as a bragging rights achievement.

Milan did the one thing that he could and that was morally right. He just got officially recognized what was a fact already.

Despite that, I gladly continue working towards my first composing title bragging rights. But then, I play video games for achievements also. What do I know?
In the end, that is a reminder that I did everything worthwhile in a game. Or did something worthwhile while composing. The title is more for me than for other people looking at it.
(Read Only)pid=21165
(14) Posted by Joose Norri [Wednesday, May 19, 2021 14:10]

Well, what else could Milan have done. If the problem was worth publishing, which it clearly was, then he could hardly publish it under his own name, not even as a "Milan, after Matti". I think they spent an evening at the Sarajevo 1984 congress on the problem.
(Read Only)pid=21166
(15) Posted by Andrew Buchanan [Wednesday, May 19, 2021 17:42]

The entire Codex Article 23 is entirely unsatisfactory. It's just a vague waft of the hand in the general direction of the topic, with no real clue. For example I was shocked to learn just in the chat here that a problem in the Codex as "P, version Q" could earn no points for the composer. The policy needs to be made completely transparent, so that new composers who weren't born when the original Codex was written can say: "OK, very clear, thanks!"
(1) What are all the Categories?
(2) Are there practical differences between versions and corrections? I see corrected versions sometimes too: does that matter?
(3) What the the requirements for each?
(4) What are the general implications for prizes etc for each?
(5) Who classifies any problem? P? Q? Album committee? What if one or more are deceased?
(6) How can we stay kinds and human in all of this? Where's the flex points?

*After* all that has all been made crystal clear, each database manager can decide how they represent this. The only point that I find myself really caring about a lot is that *every* name involved must selectable and searchable from a curated & controlled list of authors (where new names can still be added freely). Particularly given the wide variation in spellings and alphabets, it seems little enough to offer a versioner or corrector for their effort (and they're probably not going to get anything more) that their name not be lobbed into an uncontrolled and probably unsearchable comment string.

More sophisticated data modeling would allow for composers to have different *roles* with respect to any problem. However in PDB we are not there yet, and the essential of ensuring versioners' and correctors' names come from the Authors Table seems to me the best compromise. I never try to "big up" a versioner's contribution. I always annotate the "After" field with a standard phrasing, to indicate the true relationship.

For some reason Viktoras thinks "the"? question in all this mess, is that we must prevent this very reasonable current compromise. I fail to find a perspective from which this would be correct.
(Read Only)pid=21167
(16) Posted by Michel Caillaud [Thursday, Jun 10, 2021 21:59]

Sorry. Certainly Milan told the complete story, probably in a discussion or a lecture in a Belgrade meeting, but my wicked memory kept only the half-point part...

For more on H9, the booklet by Nikita Plaksin and Andrey Kornilov on the retro problems by Aleksey Troitki can be found at
The book is from 1997, but the online edition has a 2011 addendum, including the correction. Even with rough knowledge of russian language, the publicaton data can be compared.
(Read Only)pid=21247

No more posts

MatPlus.Net Forum General Codex Article 23 on correcting & versioning Prior Art