Karsten wrote:
Sorry, Rick -- the error is discussing the point at all?Nope. I guess I'll have to back up a bit.
On many technical topics, of which this is a fine example, one's aim in writing a piece is to achieve clarity. You want the reader to emerge understanding the issues. To manage this, it helps to both cover necessary material concisely and in logical order (to create a clear conceptual understanding), but also to dispel misconceptions that commonly stand in people's way -- ones that you already know people stumble over and never get past.
In this case, comparing "apt vs. rpm" happens to be among the all-time classics in the latter department.
Of course it should be discussed -- in the sense of warning readers parenthetically that "The comparisons you see frequently elsewhere of 'apt vs. rpm' miss the boat by comparing applications from two very different categories -- like comparing Chevrolets against carburetors." Elsewhere in the essay, you'd have explained that /usr/bin/apt is a package-acquisition and dependency-resolving tool, capable of dealing with either .deb or .rpm packages, and that /usr/bin/rpm and /usr/bin/dpkg are package insertion/removal tools for .rpm and .deb packages, respectively. You might have stack diagrams like these:
\n gnome-apt ---------\n aptitude | Package selection\n Corel Update |\n Storm Package Manager or\n dselect |\n console-apt ------- |\n |\n | calls\n v\n apt-get Dependency-resolution,\n | package-retrieval\n |\n | calls\n v\n dpkg Package installation & removal,\n configuration\n
...and...
\n gnome-apt ---------\n aptitude | Package selection\n Corel Update |\n Storm Package Manager or\n dselect |\n console-apt ------- |\n |\n | calls\n v\n apt-get Dependency-resolution,\n | package-retrieval\n |\n | calls\n v\n rpm Package installation & removal,\n configuration\n
If, on the other hand, a discussion bearing largely on Debian package-handling mentions the erroneous concept of "apt vs. rpm" numerous times as if it were real, and nowhere clarifies the confusion, then the odds against the reader walking away with a clear understanding given the likelihood of his swallowing whole a basic category error are rather long.
Now, you impliedly ask what is a constructive suggestion, while Peter suggests I'm "perfectly free" to fix it.
Constructive action would seem to logically require excising the erroneous drivel comparison, clarifying the levels on which the various tools work, and specifically dispelling the common misconception, somewhere in the text. If I were to try to implement that as Peter says I'm "perfectly free" to do, I'd probably end up rewriting a bunch of text, much of which has other people's names attached as attribution. Irrespective of the "wiki way" ethic, it seems downright wrong to put words in other people's mouths that way.
The ethical concern is at least a little similar to what I said yesterday to a fellow who said I should update Jahn Rentmeister's 1996 essay that I keep mirrored at [link|http://linuxmafia.com/~rick/opti.html|http://linuxmafia.com/~rick/opti.html]:
> It is funny and very good. Unfortunately, I am afraid that it is also a bit
> long and constant references to Netscape are no longer relevant.
>
> It would be nice to have up-to-date version....
Hi, Petr. I've thought of doing that, but it would have to be a
completely new essay (at least, entirely in my own words) because I
don't have the author's permission to fool with his text.
Actually, quite a few years ago, when I saved the essay from oblivion
after it disappeared from its original site, I wrote the author to
advise him that I'd re-posted it, making a few minor fixes to his
English usage, punctuation, and grammar (along with fixing a few broken
URLs), impliedly asking his permission after the fact. I told him I
really admired the essay, just wanted to make sure it didn't disappear
completely, and hoped he wasn't offended. The mail _did_ get delivered
to his then-current mailbox, but he never replied.
Anyway, without his explicit permission, I wouldn't want to drastically
revise his text: That would be doubly wrong, in that it would (further)
violate his copyright, and also attribute to him words he didn't say.
Stripping his name and substituting mine would fix the second problem
but make the first one worse.
So, the only practical and ethical way to update the essay would be for
me (or the original author, but he doesn't appear interested) to write
an entirely new one.
And I think the essay as written has a certain charm: Although the
examples and product names are dated, he uses them to make the essential
point quite beautifully -- and that point isn't dated at all.
Nautilus blows
