Re: Anatomy of Insanity?
Posted: Sep 30, 2004 9:53 AM
> Sorry. I was starting out with a bit of humor, and it
> seems as though you took it seriously. I gotta work on
> this humor without smiley's thing...
No worries. I have that same problem. :-)
> The words "technology" and "need" have never belonged in
> the same sentence together. The core of any technology
> question is the value proposition that it represents, and
> any value proposition is inherently subjective. This is
> especially true at what we would call the "leading edge"
> of technological change. Now, I think I can provide
> compelling arguments that these statements are true, but
> that would take an essay alone. Moreover, I don't think
> they're particularly controversial, so let's just assume
> that this is all about subjective notions of value. OK?
Bah. Is clean drinking water a subjective value? Maybe from the perspective of an alien anthropologist studying us. Is protection from predators (human and otherwise) a subjective value? How subjective are the benefits of the rule of law (be it de facto or de jure)?
Your response gets directly to my point. The arguments that you're making miss the real points while attempting to deal with them in a complicated manner. If I was a jerk, I might just thank you for proving my point and be done with it but I'm not so I'll continue with my rebuttal. ;-)
> > However, complicatedness, in this use, has the
> > implication
> > that the complexity (of whatever type) wasn't actually
> > addressed. Basically the feeling is that a band-aid
> > was put on a band-aid and the gangrene has been left to
> > fester.
> I understand the source of that impression, but that was
> the point of the caution in my earlier remarks. There are
> some details I left out of the account that, to a
> professional, would have greatly reduced that impression.
I guess that I have to say "yes and no" on that. I understand that you could share more mitigating information. The fact is that regardless of the circumstances, it still took y'all a bloody long time and effort to "really fix" the problem and you had a number of psuedo-fixes in the interim (which just increased the frustration and eroded the trust of the users suffering from the problem) and there's not much confidence that this sort of thing won't happen again (for example, for all of the "process" knowledge that MS has access to internally, what have you changed (like Michael's question about automated unit testing) which will fundamentally help?).
> Nor, for that matter, was my point about published works
> on software engineering intended to be a defense of either
> me or Microsoft. You are absolutely correct in saying
> that my aging analogy glosses over the distinction between
> choice and the lack of choice. On the other hand, if you
> want to say that the "complicatedness" is the result of a
> pathological issue at the organizational level, then you
> do need to show how particular decsions should have been
> made differently. The books I noted are very
> representative of the collective, organizational knowledge
> Microsoft has of the development process. While not being
> conclusive evidence in and of themselves, they do lead to
> the presumption that, at every decision point along the
> way, the best possible decision was made in terms of the
> trade-offs involved in any value proposition.
Hmm... You may think that that's the presumption that is (and/or should be) taken away but the proof really is in the pudding. Every user of MS Office has to deal with the complexity and complicatedness of MS Office everytime that they use it. Presumptions of technological prowess based upon anything but the actual products/services (and how they change (or not) over time) is basically just excuse making (and this is certainly endemic in our industry).
> This is what I keep coming back to. I can't find any
> decision that we should have made differently in terms of
> the overall value proposition. Certainly, one can
> choose to completely rewrite Word at its core in
> order to fix the problem I outlined in "Anatomy of a
> Software Bug." But, does that choice make sense in terms
> of the overall value proposition? The answer is that it
> doesn't, because the redesign would result in a different
> set of problems that would lower the overall value of the
That depends on how you calculate "value". I.e., if you don't care about how much time is wasted by the users of MS Office struggling through its complicatedness....
Also, that depends on whether or not a rewrite actually addressed the complexity and complicatedness problems. Hasn't MS and the MS Word team learned an amazing amount in the last 15 years that would help it create something more powerful, much less complicated, and arguably less complex?
> I should point out that this "value proposition" I keep
> talking about is always measured in terms of the
> customers. If we thought rewriting Word would result in
> something more people would want to use, we'd do it
> without hesitation even though the development cost would
> be high. We'd recoup that cost in the first month that
> the product would be available.
Always but not only. I.e., does security really matter? Does robustness really matter? Does usability really matter? Those don't necessarily directly drive increased users, revenue, or profits but they certainly affect them indirectly. Ah, so, then all of the attempts at quantifying their value is subjective (to y'all on the inside) even though they have objective (and subjective) effects on the users.
> > (C) The proof is in the pudding.
> Yes, it is. And Word currenly owns the word processing
> market on both the Windows and Macintosh platforms. I
> don't mean to be arrogant about this. I say this because
> we are, fundamentally, talking about a value proposition.
> That Word owns the market is pretty strong evidence that
> t Word's value proposition is pretty high.
> And, yes, there are network effects. Arguments about them
> aren't all that compelling. Any network effect argument
> one can make about Word is also applicable to WordPerfect.
> Compatibility isn't the hard nut to crack, even though
> people cite it all the time. The really hard nut to crack
> is the value proposition.
Sorry, you can't have it both ways. Those network effects and the monopoly power of the DOS/Windows platform played a large part in getting MS Word/Office into the dominant market position that it currently enjoys. Yes, of course there were other factors (such as WordPerfect's arrogance and incompetence :-).
Basically, how many MS Word users *love* it? How many people only use it because they are forced to because everybody they interact with use it? How many users curse it regularly?
Less subjectively... How many tens of millions of hours are wasted by MS Word users every year due to bugs? How many hundreds of millions of hours every year are wasted every year by the users having to (try to) deal with MS Word's complexity and complicatedness?
> So, if I sound resigned to the inevitability of software
> complexity, it's not because I think it's not possible to
> reduce software complexity in a very practical, as opposed
> to purely theoretical, sense. It's because I've yet to
> find any reason to believe that reducing software
> compexity for the sake of reducing software complexity
> actually adds to the overall value proposition. This, I
> think, was Brooks' point in the "No Silver Bullet" essay.
I don't recall any of us in this discussion talking about anything that was focused on anything but practically dealing with complexity (and complicatedness). My big abstract question revolved around suggesting that people note how the factors ("spirit" if you will) of an organization is made manifest in it's processes, products, and services.
It sounds like you have an underlying fatalism. I humbly suggest that you seriously reevaluate the distinction between complexity and complicatedness w.r.t. real simplicity.