One Book Or Two

I'm comfortable with one or two books. If AW has a position, say more sales vs. saturating the market, it would be wise to factor in their reasoning. Beyond that I'm for keeping the decision making as close as possible to the people doing the work. -- WardCunningham

Given the number of papers, it is pretty obvious we will need two books. We might even need three. The question is whether it makes sense to divide the papers by where they were presented.

I claim that from the point of view of the readers, it does not make sense to divide them by where they were presented. Instead, papers should be divided by topic. Otherwise, we will end up with a bunch of distributed programming papers in both, a bunch of frameworks papers in both, and so on. It is better to divide the papers to maximize the value of each book to a particular kind of person. This would increase bookstore sales.

I realize that each book would still contain several topics. But this is better than each one containing a dozen topics! -- RalphJohnson

Two (or three) books is fine by me, but I strongly support doing them by topic if we can. In particular I would like no visible differences on the basis of country of author or continent of presentation. I think many, both American and European, would look first to the American book if there were a distinction. This is distress of course ("we/they are better") and mixed books will help to contradict it. Later when we have PLoPs in India and Singapore this will be even more important. -- ?BruceAnderson

Agreed. We had trouble coming up with a reasonable partitioning for PLoPD 2. And I'm not entirely satisfied with the bimodal result. The first four parts have broad themes: idioms, general-purpose, special-purpose, and architectural patterns. The last four parts are domain-focused: process/organization, exposition, concurrent/distributed, and reactive systems patterns. You might argue that we've already published two books; they just happen to be bound together! -- JohnVlissides

I think a good answer might come from thinking from the buyer's perspective. If I have limited funds, how do I decide which book to buy. Is a division by location presented a division that will help me buy just the book that fits my interests? I don't think so. Organization by topics, identified by likely buyers makes sense to me.

Or we could begin to get tougher on what we consider acceptable for publication. How do we raise our level of acceptance in such a way that our acceptance criteria grow with the maturity of patterns. How do we communicate our raised expectations to the Writer Workshop Authors in ways that they will understand if they are not accepted for publication? -- NormKerth

Cope and I talked last night and I remembered some of the stuff we talked about at the end of the last PLoP. I thought recalling it here might help focus the debate.

We talked about three purposes for publishing the PLoP proceedings:

1. Provide exemplars for new pattern writers
2. Contribute to the literature of programming
3. Build the pattern community

1. Be much more choosy and publish a single volume
2. Divide the papers by topic in some way that won't be clear until after the conferences
3. Divide the papers by geography

Ok, it seems settled that there will be (at least) two books. The remaining question seems to be Book by Topic vs. Book by Conference.

Book by Topic makes the patterns community accessible as a whole and focusses on the results, the actual patterns. The conferences are hidden in the background.

Book by Conference groups the patterns by conference and might be a starting point for establishing different conference traditions, perhaps different cultures in the long run like the difference between OOPSLA and ECOOP.

I agree that dividing the books by topic helps the readership to get more out of the books. So, if we were considering two very similar conferences, there would be no problem to do it Book by Topic.

The two different conferences might help to establish two different traditions. However, I think it is much too early to speculate on this. We might get a first impression of this only after EuroPLoP.

Right now I think that we should go for two Books by Topic. There might be a (slight) chance for revising this after having experienced EuroPLoP and PLoP, and so we should plan for both possibilities. -- DirkRiehle

I'd like to remind everyone that there iw at least one more forum producing and (hopefully) publishing patterns this year, which will be the TelePLoP event at OOPSLA. Should it be part of the same milieu? I was hoping they'd publish their patterns in the same series (this is of course up to them), partly because it's a path of least resistance and provides consistency to the reading public.

(I'd be happy to have them do it. Is anyone talking to the organizers? What about the other patterns events at OOPSLA? What happened to the papers on distributed programming patterns from the workshop at last OOPSLA? -- RalphJohnson)

But it seems a burden to combine them into the same editorial board.

My own opinion is that the editorial boards should communicate with each other, and maybe even have joint membership, but that they should be separated by function (EuroPLoP vs PLoP vs TelePLoP) rather than by topic (though TelePLoP will be functionally distinct). In two or three years, we can convene a separate editorial process to establish an index for all volumes, to tie them together.

(My opinion is that it is high time to start trying to organize things. It will take a long time to do that, and it is not too early to get started. Cope is right that we don't know how to partition things, and that we don't know how things will evolve. We can guess, but that is all. -- RalphJohnson)

Besides, it's hard to establish the "right" topical partitioning at this point -- it's been hard enough for us to delimit chapters in the two volumes we have! Think of it this way: do we know, at this point, what the topical division would be? What are the two topics? How would the series evolve? If we have a "business patterns" volume, will there be a "business patterns" volume two? Will it transcend the conferences we foresee for the future? What are the plans for orchestrating this in the future? -- JimCoplien

No, we don't know what the right partitioning is. I don't think that plopd-1 and plopd-2 have provided us with the final answer, nor will be able to do it this year. However, once that we are aware of this problem, *trying"* to work out classification schemes, as problematic as they might be today, might provide us with the experience to tackle this problem later on.

I think it is important to show that current proceedings structures are not the final word--a fine thing to say in the introducing editorial--but at least we might get the experience to come up with a satisfying classification scheme in some years. Satisfying relative in time, of course.

Thus, I don't see negative consequences with a Book by Topic approach, but only experience to be gained. I don't think we will determine the future by our current work.

This is the first year that there will be more than a single volume. We clearly have the right to experiment. It is obvious now that the prefered solution is to do two books by topic, so we have to discuss amongst this year's editors how such a partitioning can be done the best way.

I feel unsure about TelePLoP. OOPSLA is a long time ahead, and by the time TelePLoP takes place, authors should have provided already with their final submissions. How do the TelePLoP organizers see their work with respect to the PLoP and EuroPLoP patterns? -- DirkRiehle

(But, actually, TelePLoP and PLoP are closer together than EuroPLoP and PLoP, right? -- JimCoplien)

In time, yes. But with my present understanding, PLoP and EuroPLoP have been coordinated, in particular they have a common proceedings revision deadline. It is the deadline which has been announced and which ties PLoP and EuroPLoP together. If TelePLoP should be part of this, we would have to move the deadline into November or December.

Well, we could do this, but do we want this? One of the arguments I saw against joint books by topic is the time gap between EuroPLoP and the revisions deadline. I feel rather uncomfortable with this and if we continue books by topic it will become more difficult to separate PLoP and EuroPLoP in time (which would be a fine thing IMO, but interferes with a joint revision deadline). -- DirkRiehle

I can see Dirk's concern. November/December is a long time from EuroPLoP. However, we intend to limit attendance at TelePLoP to a dozen or so. With such a small number, we should be able to turn them around faster. It also gives us more flexibility; for example, we could do some screening and more extensive shepherding before TelePLoP itself, in order to prepare the patterns for publication. It also seems possible to maintain the PLoP/EuroPLoP deadline (is it the middle of October?) and allow the small number of TelePLoP papers two weeks later.

But I would like a bigger question answered. Do we want to include the TelePLoP patterns in the books? If so, we will work the logistics and make it happen. -- NeilHarrison

(From my point of view, you're welcome to team up with us. I assume that the number of good patterns from TelePLoP will not constitute a volume of its own? If so, and if everyone agrees, we would have to discuss again how to distribute work. -- DirkRiehle)

It seems to me that there will be more and more PLoP conferences each year. In just three years the number of conferences had doubled! I expect this trend to continue. It would be impractical, I think, for each such conference to generate its own book. The volume would be too high, and the quality would suffer.

I think one book per year would be best. "The best of the PLoPs of year XX". I doubt that any other course is realistic. As such, perhaps we should follow that course even this year. -- RobertMartin

(I don't know what the final quality of the papers will be in October this year. If it is that bad, we might indeed do only a single volume--it surely would reduce our workload. Right now, it looks like we have to expect two books. In the long run, you might be right, but I cannot predict this. -- DirkRiehle)

(The decision about one or two books, about an organization by conference or topic, we need to wait for the PLoP/TelePLoP submissions, and for the results of the conferences. -- FrankBuschmann)