Home | Login | Recent Changes | Search | All Pages | Help

WiseGuyMusing

About "perfect session scheduling"

I've had several recent exposures to attempts at prefect session scheduling. Haven't had the opportunity to say what was really on my mind - the myriad ways the exercise was just silly. Following is in terms of potentially scheduling sessions for AYE, 2002, but isn't really about that. It's about these other experiences, mostly. I still have some of the goo on me.

A better schedule for AYE 2002:

Well . . . The list of aspirations in your Johanna's of 2002.9.26 is unresolvable with current information. I'm suspicious that the whole list is unresolvable with any amount of information. Add in the other points of view expressed, and it becomes quite the fur ball. It seems to boil down to "the maximum good for the maximum number of people" in the somewhat restricted domain of session selection and attendance at AYE 2002. Seems like a hard problem to me. Religions, governments, and theories of the same have been chewing on that one for several thousand years. Might be a bit of a challenge to solve this one in a month or two. But that's just me.

At any rate, even if we could construct an effective optimization / tradeoff for the several goals listed, for you, your tradeoff matrix would undoubtedly be different from that of other people. Your tradeoff matrix on behalf of other people would be different yet again (assuming the list of things to optimize for stays the same - a silly assumption really), and likely wrong anyway. Of course as the conversation proceeds, any optimization in terms of "the greatest good for the greatest number" has to take into account the conversation to date, and how anyone might be feeling about that.

I'm not sure I could even model the boundaries of the combinatorical explosion for this problem. That boundary leaves out other takes on the problem:

• People like Jerry, who might decide to throw some chaos on the situation if it looked like it was settling down.
• My preference which might be to pick some sufficient solution, any sufficient solution, and get on with having a good time. Unless the wrangle around the intractable problem is sufficiently amusing, in which case watching that counts as a good time. (Sorry. I've been exposed to a fair amount of "everybody worrying about everybody else" over the last couple years. I've still got some of the goo on me.)
• Some folks who will need to show off their analytical ability by coming up with the ideal optimization scheme. If we've got two of them, one of them will be unhappy.
• Then there's the people who know what should be important to the other people involved, so will insist that we ought to optimize for what they should want. There's a bonus in there, in that the miscreants are being fixed, as part of the bargain.
• etc.

This is "a hard problem" on several levels.

There are at least as many workable, sufficient solutions to "avoiding unattended sessions" as there are people planning to attend AYE. Pick one and be done. Of course, if we replace insufficiently subscribed sessions with other sessions, pretty soon everybody's conflicted. There are only "good" sessions left, so everybody wants to be at every session. (This assumes that signup for sessions expresses a preference, not some martyr style sacrifice. Also that there's some similarity in the attendee's idea of what is a "good" session.) Is making everyone conflicted about their session attendance a net gain for the people involved? A net gain in happiness? Are we going for minimization of misery?

• One solution that eliminates any potential conflict of aspirations is if everyone wants exactly same thing at exactly the same time. I'm not sure how I'd know that when I saw it, let alone predict it ahead of time for an internationally dispersed group. Maybe some kind of unadulterated groupthink would get us there. Is that an OK solution?
• Maybe we want one session, presented a bunch of times. Or one presenter. Or one attendee. Then we're free of choices once again. Seems kind of boring, but that's not on the list of aspirations.
• Maybe we want all crappy sessions all the time so nobody is bothered by missing one. Also only crappy hosts, and crappy attendees for the same reason? Would certainly minimize disappointment.
• If we're concerned about not hearing that people maybe also wanted to be somewhere else, we could prevent them from talking, or just not listen. Duct tape and earplugs are one approach.

I'm for continuing a content-free scheduling discussion, perhaps indefinitely, hopefully building to acrimony and frustration. I'm having fun being a wise-guy, and, of course, it's all about me.

As for attending sessions, I'm hereby up for bid. I'll guarantee to attend, or not to attend, whatever combination of sessions pays me the most. For a nominal fee, I could be persuaded to stay at home entirely. So come up with a preferred schedule for jb, with accompanying bid. Cash or money order. No COD. Optimizing that offer is left as an exercise. It's more tractable than this one.

- JimBullock, 2002.9.27

So, we make the best plan we can and adapt as the days play out. Last year the juggling didn't seem to impact anyone's enjoyment of the conference. I doubt that any design or plan will ultimately change that.

But, I am enjoyiing the musings.

- BeckyWinant 9-26-02

Well, if you enjoyed the musings that makes at least two of us. On rereading, it drags a little in a couple of places, but could easily be something with a little work. I'm glad someone "got it." I was a little worried that this bit wouldn't be taken with tongue in cheek. There's no standard wiki markup for that.

As for requirements, I've been playing with uncertainty in this and other areas for a while. I think, in all seriousness, that a perhaps more useful point of view goes like this:

You can know a set of requirements ahead of time, which can be well formed by some important, useful criteria: SMART, unambiguous, consistent, blah, blah, blah. But once you've got the system in hand, the new system, of developers, users and existing vs. imagined system is a different system. The requirements you'll develop, or the way you will understand them will be different. This is most obvious in the requirements that show up in testing. Ahem:

The rich interaction of engaged humans with the system itself cannot help but generate candidate features and refinements that were at a minimum impractical to attempt to know ahead of time. The experience of creating the system, and the presence of the system itself have added information. So count on "new" requirements coming up at that time. The inescapable fact that "the requirements" as known at the beginning are different from, usually less than "the requirements" as known at the end merely shows that you've learned.

Operationally, for development management, I've found a useful definition of success is: "Meets the intended requirements, while stimulating lots of additional ideas." You can't give up on shipping what you said you would, merely because you've thought of something new. And you can't rule out discovery of some new feature without doing violence to the very same creative intelligence you must employ to get stuff built.

The other effect going on is that having the system in hand changes the price and risk points making many candidate requirements feasible that weren't before. (and some others impractical as well.)

There's a draft of an article length piece behind this thought, which I'm not going to get to any time soon, unfortunately.

- JimBullock 2002.9.27

In Ron Jeffries' words, "The fact that you know more today, and are more capable today, is good news about today, not bad news about yesterday." -- LaurentBossavit (tomorrow)

Ah well. Now I've got a new thought experiment. "Is there a way to combine this bunch of people in a comfortable environment, and still have a bad time?"

I can make a hard problem out of anything. (That's a feature, right?)

JimBullock, 2002.09.29

I'm not sure that exploring a bad time is "time well spent". Maybe there's a thread for pessimists who what to attend a conference to discover everything they need to avoid. Boring for the optimists. Depressing for those who seek a balanced view.

As for Johanna's : Of course, for the non-valuable conferences, just like the non-valuable software, it doesn't matter when you schedule what, because it's useless. I'd *much* rather have the problem of too much useful stuff than not enough useful stuff. Absolutely! Get the topics and sessions that matter to people.

I guess we need a sense of what matters... and we find about how much of what we offer matters and that titles and short descriptions are not enough, since people will change their mind 5 minutes before the next session.

Having too many good sessions is a good thing. Comflicts about what session to attend is not a bad thing. Now, back to Jerry' s problem - how do we get this across to people early in the cycle?

- BeckyWinant 9-30-02

We could try the standard educational system ploy: grade people on how nicely they stand quietly in lines. (After all, nothing is more important than waiting in conforming style.)

Since that fails, your on your own for bad karma with such a crew. Nutz to herding cats!

-BobLee 2002.09.30

Sign up for the ones that blaze out at me as important, get there, meet people, and re-assess. I won't know which opportunities will appeal the most until I know the people involved. Which workshops will challenge me? Which will push me out of my comfortable box?

I am really excited to see where I land...

-MaigWorel 2002.10.1

--BobLee Oct 4, 2002