Search engine companies have traditionally
avoided indexing pages with a low number of
incoming links. But that policy is changing
as search engine companies seek to increase
the size of their index.
The point being that personal sites are
going to start mattering more than they
have in the past.
Let's try that again. If you take a large sample
of people, and you generate random numbers (0-9)
and you ask them to repeat the list of numbers
just read, then you find that beyond 6-7 numbers
people can't do it. That's why the phone company
chose 7 digits. It's not true for everyone,
but that limit does fall sharply off around there.
(And it takes about 30 seconds before your short-term
memory starts to loose track.)
If you play 7 tones for a person, then play
one of those at random, they can recall the
number for the tone. However, if you go
beyond 7 choices, they can no longer distinguish
the random tone.
I'm in california where the license plates
are limited to 7 digits and letters. You'll
find lots of similar examples.
So, the point (which I didn't make clear
enough) is that we (humans) start to loose
track after six, and "things"
begin to feel cluttered. That's why I was
suggesting that more than 6 folders for email is
overkill.
You were right about my point being a weak one.
I was really only referring to short term
memory. You could certainly remember hundreds
of email folders. My reasoning is more fuzzy
than logical, and more empirical than provable
(but then again aren't we?)
Ironically, I can't recall where I first
read about this phenomenon, but here's a
Fun test.
PS. About the telephone lookup. The one feature
you need is to support a multi-line entry,
then you can use it as a general index-card sort of thing...perl good
Humans have the ability to easily distinguish
among 6-7 things. Past that many it's inherently
difficult to distinguish in short-term memory.
When the US phone companies first developed
the 7 digit phone number, it was because this
many digits is around our natural memory limit.
So, for email, I keep folders to a minimum of
about 6. But because time is so important, I
tier those, so that anything older than a week
goes into a mirror folder structure under OLD.
Then anything older than a month is moved into
ARCHIVE. And the archive stuff is compressed,
so I have to really want to look at it.
inbox root etc OLD/inbox ARCHIVE/inbox.tgz
Loose information is another problem, with
a simple index-card like solution.
A lot of the information we need is small, like
"joe's phone number", and doesn't warrant a whole
file. For that I actually throw all the information into a single big file, where each
datum is one line (grep-able). The information
has no structure. I often cut and paste random
stuff. Then I have a search that just pulls out
entries which match all terms:
For shared stuff, I really like having an unofficial document system, and my
favorite is The Moin Moin Wiki because it's fairly simple to use and
install.
I wouldn't mind ignoring the setuid and locking
issues entirely. I'd love to have a samba
server under windows to provide "programmed"
filesystems. It's very difficult to write a
low-level windows FS, whereas a lot of people
would like an open-source user-land server.
The reason is not even for providing files,
but for providing programmed interfaces
which look like files to the system. Eg.
I'd like to make a "registry" filesystem
which looks like a normal windows share
to the local machine.
Thanks,
These things are great, but they still seem
too complicated.
I'd like to see a
wrapper program that traps all system
calls and maintains a reversible set of
file system changes. In other words, I'd
like to be able to say the following for
any software
tar xvfz pkg.tgz
cd pkg
installer -p pkg make install
uninstaller -p pkg
and "the right thing"(tm) would happen.
installer would record the changes
being made in a way that could be
rolled back. In fact maybe installer
could produce an RPM.
Many moons ago I used scotty, when I was a sysadmin for the computer science dept. at columbia u. (about 300 unix hosts, with a bunch of routers.)
It's a tcl/tk-kind-of-thing that has several useful tools for watching SNMP, and other sorts of network monitoring (ICMP). It had the ability to create some pretty reasonable network maps and do dynamic discovery (based on watching routing tables.)
http://wwwhome.cs.utwente.nl/~schoenw/scotty/
http://www.ibr.cs.tu-bs.de/projects/scotty/
PS. For SNMP there are also perl modules that are helpful for writing monitoring scripts.
Slashdot Mirror
The point being that personal sites are going to start mattering more than they have in the past.
- It's linux and I can muck around.
- It's got WiFi.
- Setup to handle printing for my home net.
Been there done this. There must be dozens of these kind of devices on the market already.Let's try that again. If you take a large sample of people, and you generate random numbers (0-9) and you ask them to repeat the list of numbers just read, then you find that beyond 6-7 numbers people can't do it. That's why the phone company chose 7 digits. It's not true for everyone, but that limit does fall sharply off around there. (And it takes about 30 seconds before your short-term memory starts to loose track.)
If you play 7 tones for a person, then play one of those at random, they can recall the number for the tone. However, if you go beyond 7 choices, they can no longer distinguish the random tone.
I'm in california where the license plates are limited to 7 digits and letters. You'll find lots of similar examples.
So, the point (which I didn't make clear enough) is that we (humans) start to loose track after six, and "things" begin to feel cluttered. That's why I was suggesting that more than 6 folders for email is overkill.
You were right about my point being a weak one. I was really only referring to short term memory. You could certainly remember hundreds of email folders. My reasoning is more fuzzy than logical, and more empirical than provable (but then again aren't we?)
Ironically, I can't recall where I first read about this phenomenon, but here's a Fun test.
PS. About the telephone lookup. The one feature you need is to support a multi-line entry, then you can use it as a general index-card sort of thing...perl good
So, for email, I keep folders to a minimum of about 6. But because time is so important, I tier those, so that anything older than a week goes into a mirror folder structure under OLD. Then anything older than a month is moved into ARCHIVE. And the archive stuff is compressed, so I have to really want to look at it.
Loose information is another problem, with a simple index-card like solution. A lot of the information we need is small, like "joe's phone number", and doesn't warrant a whole file. For that I actually throw all the information into a single big file, where each datum is one line (grep-able). The information has no structure. I often cut and paste random stuff. Then I have a search that just pulls out entries which match all terms:For filesystems, I found the reiser guys have some very pertinent ideas, albeit in need of further development. http://www.namesys.com/whitepaper.htmlFor shared stuff, I really like having an unofficial document system, and my favorite is The Moin Moin Wiki because it's fairly simple to use and install.
The reason is not even for providing files, but for providing programmed interfaces which look like files to the system. Eg. I'd like to make a "registry" filesystem which looks like a normal windows share to the local machine. Thanks,
I'd like to see a wrapper program that traps all system calls and maintains a reversible set of file system changes. In other words, I'd like to be able to say the following for any software
cd pkg
installer -p pkg make install
uninstaller -p pkg
and "the right thing"(tm) would happen. installer would record the changes being made in a way that could be rolled back. In fact maybe installer could produce an RPM.
It's a tcl/tk-kind-of-thing that has several useful tools for watching SNMP, and other sorts of network monitoring (ICMP). It had the ability to create some pretty reasonable network maps and do dynamic discovery (based on watching routing tables.)
PS. For SNMP there are also perl modules that are helpful for writing monitoring scripts.