Re: OS questions draft, III

[Roger Dingledine <arma@mit.edu>]

In message <98784d1b.35ea22a4@aol.com>>, eamorical@aol.com writes:
>As I just logged on, AOL had a news brief on a study to be
>published this week in The American Psycologist showing
>that use of the internet is depressing. The more people use
>the internet the more depressed they get. They must not be
>involved with Linux.

Neat. How public (popular) is "american psychologist"?

>There seems to be several topics going on here. First the
>(open source, free software), GPL issue. I'm looking at a
>write-up by Matthew Thomas on RMS's New Zealand talk
>on 8/28/98. Following why GNOME it states "This applies to
>any proprietary software running under GNU/Linux -- the
>implication is that the role of WordPerfect, Oracle, Informix
>et al. on the OS will be a temporary one, until the free
>equivalents are available.". I gather the goal is to create an
>entire suite of GPL applications. Not only do I subscribe to

Yes, that is correct. That is what Debian is doing now. That's
why things that we take for granted, like ssh, most of the
X servers, mysql, etc are in the non-free section of Debian,
and down-played as much as possible (they don't arrive on the
CD). Debian's (aka GNU/Linux's) goal is to create an
environment that is good enough to use, that's created entirely
out of free software.

>this I would go further. Free information thought I don't know
>how you GPL information. In other words take human
>knowledge and set it free. Take information out of the hands
>of people and companies who horde it and sell it to you in
>dictionaries, encyclopedias, books, etc. Never in human
>history has there been such an opportunity to get back
>millions of times what you give. I hope to be able to set up a
>server to dish out free information. No advertisement except
>maybe one for Linux. This would be my contribution to

Setting up the server is not difficult, for the most part --
if you had good information, seul or debian or whoever would
be happy to include it in their site. It's much more difficult
to get the information and put it into a form that's useful.

As you get more and more information, you need more complex and
powerful means of sorting, scanning, and browsing it. Consider
if the web consisted of 50,000 text files, each of which was 10
megs of normal text. It would not functio.

[snip]
>within feet of a window. Several weeks ago I looked up from the
>computer and saw a raccoon looking in the window. I would like
>to get a digital video camera and try to record these events and
>GPL them so everybody view them and enjoy them. Anyway this

Huh? The main focus of the GPL is to force derivative works to
remain free. By putting things in the public domain (that is,
removing all licenses), people can take an item and do whatever
they want to it, including selling it or making derivative works
and selling them. The trick is that everything they distribute
has the same license, meaning that people are able to get it
including the original thingie (eg source code) used to generate
the product. Presumably there will be somebody who obtains the
product and gives it out for free, but there's nothing that
enforces that you have to distribute gratis (without charge).
For GPL, think "free speech" rather than "free beer".
Check out www.gnu.org.

Are you trying to make it so people won't sell your pictures?
If so, you don't want the GPL. Hm. A simple copyright does not, I
believe, restrict distribution of pictures; it's the license
(little notices like "copying expressly forbidden" and the like)
that restricts that.

I guess "derivative work" has some meaning though, since people can
use the gimp or photoshop to muck with the images. But I'm not
convinced this is what you had in mind...

>is the reason I am so interested in seeing the multimedia and
>other emerging technologies incorporated into Linux.
>I have been thinking about the education questions and somehow
>it seems to me that there should be a way to ask the general
>question and interpret it in a variety of ways such as education,
>business, games, etc. For example we have a general question
>about the price of software and the availability of software. Now
>if we somehow knew if the survey taker was wearing their
>educational hat or their business hat or their gaming hat, we
>would know to interpret the general software question as
>educational software for example. This goes back to primary
>user-type. But we don't have an educational user-type. Is there
>some other way to do it. Even if we determine that someone is
>going to school, we don't know if they have on their educational
>hat, gaming hat or some other hat. Also poeple who are not in
>school have an interest in education, parents for example. I
>believe a person is only to do the survey once, but what if they
>truly want to take the survey more than once wearing different
>hats. There is no way to stop them. Anyway more to think about.

Well, there are ways to stop them, but none is foolproof. Most
of them would be stopped by a little notice that said "please
only take this survey once."

So your concern is, "what if they have different answers to a
question depending on whether they're looking at it from the
viewpoint of gaming, or sysadmining, or being an education
user?"? I will grant you that they might want a certain feature
more for some of their activities, and not care about it for
others of their activities. But they're a single person. The
ideal case would be that they merge all of their concerns and
figure out, overall, how they view that issue. But yes, that
takes a lot more thought.

The reason why I want as few user types as possible, is so we
have as many people in each user type as possible. We really
can't start investigating our data until we have a lot of data
points for each type. If we have 1000 people take our survey
total, then we'll have perhaps 400 in the most popular type.
This is enough to actually make some statements from. If we
have 20 user types, then we'll have maybe 100 in the most
popular one, and that's pushing it. I think 'educational user'
fits pretty well under 'general user', where you need to do
things like word process or play games.

What you're getting at, it seems, is that for a given question,
you want to know *why* they said that issue was important to them,
and more importantly what they thought the question was asking.
Is this a better rephrasing?
That's a very difficult thing to pull out of a surveyee in a
reasonable amount of time and energy. :)

>Bob

--Roger