Beyond Writers Workshops

I've long suspected that the WritersWorkshops which are held at the various PatternConferences? throughout the year aren't doing enough in helping patterns authors create literature. There are way too many examples of poorly written patterns and pattern languages that have been "workshopped" -- some have even been workshopped numerous times and still come up incredibly lacking.

This does not mean to say that WritersWorkshops aren't good. It only means that we need to understand what isn't working and look for a solution.

If you've ever experienced a WritersWorkshop, you may know that not a lot of time is given to each author. It simply isn't possible to give an author more than say, an hour, during a three-day conference, in which that author's work will be reviewed. This is a problem. One hour just doesn't cut it for real, well-thought-out feedback, especially when conference attendees have to read/review 5-8 other papers during the conference proceedings.

But that problems is compounded by yet another problem. Consider who is present in these WritersWorkshops, doing the reviewing? Who is actually there, offering feedback to the author? Are they the author's ultimate audience? No, they are typically fellow patterns authors, who have written on a similar topic, and are at the conference, in the same workshop, because they've submitted a paper on the same topic.

So, what's wrong with that? Well, patterns and pattern languages are meant to be accessible to, and used by, laypersons. But none of the individuals in a typical WritersWorkshop are laypersons. They are usually experts in their fields. So although they can help improve a work, they cannot really offer good feedback on whether or not the writing is working -- i.e. making sense to the layperson: the ultimate consumer of the goods.

I believe the problem stems from the very roots of WritersWorkshops. These roots come from the creative writing community. I'd like to argue that we are a different kind of community. We are not writing poetry or fiction - we are writing a new, practical literature that laypersons will be able to use to build real things, using master knowledge.

So now we are moving into a favorite topic of mine, called using patterns. The question is, who is actively reading patterns and patterns languages for the purpose of educating themselves to use their new knowledge to do something? The answer is, many of the folks who regularly attend patterns study group.

In my experience, these individuals are perfectly suited to the task of providing authors with feedback. Study group participants are usually bright, well-read individuals. Many of them are laypersons, and some have years of experience in a field, but lack experience in certain areas.

In addition, because our community now has so many patterns groups, authors may give their drafts to several groups, distributed world-wide, in order to get back a large, and thorough body of feedback.

Each group which gets to review a work, would study it before meeting, then have a dialogue on the work, in which feedback is captured, followed by giving the author the feedback either in a document or within some online forum.

I believe that this is the next step, beyond WritersWorkshops.

I'm currently experimenting with this to see what kind of feedback I get after giving a draft of my own pattern language to several patterns groups. If this proves to be successful, perhaps we can move closer towards creating a new, valuable body of patterns literature by making this established practice. -- JoshuaKerievsky

Actually, Joshua, I think we're already doing that. For instance, when we were writing the DesignPatternsSmalltalkCompanion our patterns were workshopped by two different reading groups -- the group in Urbana led by RalphJohnson and a reading group at KSC. We found this feedback to be invaluable in making the book better. Ralph's group in particular has workshopped many books -- like the new one by MartinFowler on refactoring, and many others. I think that group of people may be the best thing to happen to patterns since WardAndKent first got together.

-- KyleBrown

I'm talking about something a little more formal, Kyle. Yes, some people have submitted works to groups before (and I personally spent many hour and days reviewing two drafts of Martin's new book). But unlike WritersWorkshops, this is not a formal process: it isn't done for the vast majority of patterns writings. I'd like to see pattern languages register for and do globe trots before they are ever published. This would mean that study groups would have to be interested and able to put in the time to help authors. My guess is that if the content of these drafts is worthwhile, groups will participate to help improve them. But the patterns community should get a real boost if this process is formalized, so it becomes accepted practice to have every would-be piece of literature visit several groups before it is published. -- JoshuaKerievsky

Joshua, so how do you think this would work? I know that our patterns group in NC would be interested in participating if it was easy enough. It seems like we'd need at least the following:

A place for Authors to register their work (wiki?)
A "standard" signoff from the publishing companies (like AWL and Wiley)
A "standard" mechanism for feedback -- either more Wiki pages, or a more traditional forms-based web approach.

The benefit to the patterns groups is that they would get the latest work early. The benefit to the author and the publisher is that their work is reviewed better. I suspect that AWL and Wiley would both be interested in something like this. John? Any comments from the publishing peanut gallery? -- KyleBrown

Yes, we need (1) and Wiki seems to be a fine place, though I'd say authors should also echo the registration of a new work on a patterns mailing list, so people learn about the existence of some new work, without getting access to it.

We also need a list of the study groups which will commit to giving feedback. This should not be too difficult, particularly if we use Wiki.

Getting feedback is tricky. I've experimented with this for a while in the study groups I've participated in. After some time, I finally hit upon something I call a DistributedDiary?. This is a useful way to get quality feedback and is documented in my pattern language on study groups.

I'm curious about (3) - your idea for a form. I think this has great potential. I'd favor a form that would be partially static, and partially customized by each author.

As for (2), I'm not sure the publishers need to be involved at this stage. Publishers want to make money. They tend not to care about literature. I think the practice we are talking about will become popular with authors, and publishers will go along for the ride. But perhaps you had something else in mind? -- JoshuaKerievsky

My argument for (2) is that the publishers have found that good books make more money than bad books. Good reviews on Amazon, in magazine articles, word-of-mouth, etc. translates to $$$$. In that respect, they should care, and, at least where Wiley and AWL are concerned, they do. They want quality product because it sells more units. --KyleBrown

That's interesting Kyle. I have a great fondness for AWL, but Wiley? That is a surprise. They seem to have published some fairly poor patterns books. Have they recently changed their tune?

I think sometimes the problem is that patterns and pattern languages are workshopped by people who are pattern-experts, but are not necessarily domain experts in the subject matter of the pattern itself. So they are able to obtain good comments on patterns form, but less so on technical substance. I see this happen frequently with non-OOD patterns. -- BradAppleton

Yup. That's the biggest single problem with the PLoP workshops. So what do you suggest we do to help this out? -- KyleBrown

How's this: when an author goes to register their new work on Wiki, a new page is created. That page will follow a template. On the page, patterns study groups will volunteer to review each work. In addition, there will be a section for expert groups - groups that have expertise in a subject, and can provide expert feedback. People on Wiki can suggest which expert groups an author can approach to obtain a thorough review. Then, we'd need to see to it that those reviews actually happen.

At the end of the process, a paper can get some logo, which will sybolize that it has be thoroughly reviewed. This will mean that it has been reviewed by X number of study groups, and Y number of expert groups, along with being traditionally workshopped at a PLoP conference. Is this raising the bar to high? I don't think so. -- JoshuaKerievsky

Very interesting idea. Maybe the HillsideGroup should be the "logoing" organization? -- KyleBrown

I'm not sure. It seems to me that the HillsideGroup took the patterns community to a certain level. But perhaps it is now time for others from this community to take patterns to the next level. If you, Brad, myself and others have interest in this particular practice, perhaps we can all work together to make this a reality. -- JoshuaKerievsky

Joshua- MassimoArnoldi and I wrote a very experimental PatternLanguage called TimeTravel for EuroPlopNinetyNine?. (http://www.argo.be/europlop/Papers/Final/Beck.pdf) The idea was to strip the patterns down to the absolute minimum and embed them in a DevelopmentNarrative?. I validated the patterns by asking a friend to use the patterns in a day-long programming exercise (the kind of thing you might have done in college). I got good feedback about what was clear and unclear from this. Now that the workshop is over, I can say I also got good feedback there, but of a very different sort. It is nearly impossible to give technical, content-oriented feedback in a WritersWorkshop. To do that you have to actually try to use the patterns. -- KentBeck

Absolutely, Kent! Two good things here: I've been throwing around the idea of doing this myself -- I'm working on a book that I THINK should have a few patterns embedded in it but it's mostly (90%) not about patterns. The notion of a narrative with patterns embedded (perhaps as sidebars) resonates really strongly with me. Also, I agree wholeheartedly that you need someone to TRY to use your pattern language to get feedback about it. We forget about "The Oregon Experiment" when we stop at "A Pattern Language". --KyleBrown

A writer's workshop at PLoP can only afford to give an hour or two of attention to a paper. The more time you can spend reviewing something, the better. That is one of the advantages of a local reading group. The local reading group might be bigger and have a lot of experienced memebers, or it might be smaller and have less experience. The quality of the reviewing will depend on the quality of the group.

An even better way to figure out how to imptove patterns is to use them. Get some experience with them. Have other people read the patterns and watch them use them. WritersWorkshops are better than having one person review a paper on their own, but they are still limited. -- RalphJohnson