<firstname>.<lastname> for UNIX accounts
gusmeister
accounts. This has two potential problems: The login name may be very long
(over twenty characters) and the dot in the middle may cause problems for
some applications.
I know that the Solaris 'admintool' does not accept that format but
'admintool' has always been broken anyway. Does anyone know of other
problems we may encounter using the <firstname>.<lastname> format.
Tony Curtis
Solaris user names need to be <= 8, see "man useradd".
If you're doing this for email and you're using sendmail,
you can set up aliases, and maybe build a genericstable:
http://www.sendmail.org/m4/features.html
(Similar functionality will exist for other MTAs)
hth
t
--
Wait up guys! I've gotta empty my grass bag.
Paul Southworth
I don't know why you're doing this, but it's usually done for the
wrong reasons. It does not create predictable login names; it makes
them less unique.
So what happens when your employee has a bunch of umlauts in his name?
Or god help you if you hire any Norwegians. So you'll just butcher
their name until it fits in the policy? And when you hire your
second James H. Smith, what then? How can someone predict which one is
which?
http://www.sendmail.org/faq/section3.html#3.5
My suggestion is:
1. Implement a directory if you don't already have one.
2. Let users pick their login names. They like it better that way
anyhow. Usually a mandatory naming scheme is resented as a
kind of bureaucratic anal-retentive tyranny by users - yes, they
actually do take it personally.
Regards,
--Paul
gusmeister
create long usernames (e.g. clearcase_albd). It may issue a warning but
that's it. 'useradd' also allows you to use a dot. It's other applications
that I'm worried about.
"Tony Curtis" <tony_c...@yahoo.com> wrote in message
news:87vga0n...@limey.hpcc.uh.edu...
Ed Wensell III
>
> My suggestion is:
>
> 1. Implement a directory if you don't already have one.
>
> 2. Let users pick their login names. They like it better that way
> anyhow. Usually a mandatory naming scheme is resented as a
> kind of bureaucratic anal-retentive tyranny by users - yes, they
> actually do take it personally.
>
Or:
3. Assign everyone a unique number/character string. Takes care of
renaming problems. Especially after marriages.
--
Ed Wensell III
NetBSD/Alpha at home - Solaris/SPARC at work - OpenVMS in a past life
E-mail address is valid if you know the appropriate bits to drop.
wibble?
Tony Curtis
[ top-posting re-arranged ]
>> On Mon, 6 May 2002 20:35:04 -0400,
>> "gusmeister" <g...@meisters.org> said:
> "Tony Curtis" <tony_c...@yahoo.com> wrote in message
>> Solaris user names need to be <= 8, see "man useradd".
> Actually, 'useradd' (unlike 'admintool') under Solaris
> 7/8 does allow you create long usernames
> (e.g. clearcase_albd). It may issue a warning but
> that's it. 'useradd' also allows you to use a dot.
> It's other applications that I'm worried about.
"A future Solaris release may refuse to accept login
and role fields that do not meet these requirements."
So it's accepted now, but things may start going horribly
wrong next time you upgrade the OS. If this is for email
aliases (you haven't confirmed this anywhere in the thread
yet though) see my first post and the other thread
members.
Kjetil Torgrim Homme
>
> We are considering using the <firstname>.<lastname> format for
> UNIX accounts. This has two potential problems: The login name may
> be very long (over twenty characters)
up to 16 characters works pretty well.
> and the dot in the middle may cause problems for some
> applications.
yes, but they are buggy if so.
> I know that the Solaris 'admintool' does not accept that format but
> 'admintool' has always been broken anyway. Does anyone know of other
> problems we may encounter using the <firstname>.<lastname> format.
beware company politics: what happens when the second Jim Clarke
appears? do you assign him "jim.clarke.2"? it's better to use
traditional usernames, since you can wing it by giving one of them
"jimc" and the other "jclarke" or whatever. make email aliases for
their full names instead (if possible).
--
Kjetil T. your computer can help researchers understand how
proteins work. http://folding.stanford.edu
Stefan Berglund
> We are considering using the <firstname>.<lastname> format for UNIX
> accounts. This has two potential problems: The login name may be very long
> (over twenty characters) and the dot in the middle may cause problems for
> some applications.
I suspect that what you really want is to have "intuitive"
emailaddresses[1], so I'm assuming this in my answer[2]...
Third: Which emailaddress do the next person get that has the same name
as an existing user?
What happens with the address of codeslave John Doe, when Big Boss John
Doe joins the firm?
Will the one that has john...@company.com be happy when he gets the
email for john...@company.com, and vice versa?
This gives you more problems than you solve.
[1] The argument usually is that "you know peoples names, so it would be
easier if they had the same mailaddress as they are called".
The problem is that the logic is trivially wrong as it is not
possible to have a one-to-one map of a nonuniq namespace (peoples
names) to a uniq namespace (peoples loginnames, emailaddresses or
whatever) without introducing a "uniqifying" element that is
inherently nonintuitive.
[2] But as [1] implies the reasoning is the same for login names as well.
--
/Stefan
sbl+...@dd.chalmers.se
Life - the ultimate practical joke
Alan Coopersmith
|Does anyone know of other
|problems we may encounter using the <firstname>.<lastname> format.
Increased user hatred of sysadmins and higher rates of carpal tunnel,
especially among users forced to type very long names to login?
Breaking the column alignment of ps, ls -l, and just about anything else
that lists usernames as one column in a report.
Doing this as e-mail aliases is one thing, but doing it for usernames
just seems like BOFH behavior.
--
________________________________________________________________________
Alan Coopersmith al...@alum.calberkeley.org
http://soar.Berkeley.EDU/~alanc/ aka: Alan.Coo...@Sun.COM
Working for, but definitely not speaking for, Sun Microsystems, Inc.
Chris Watkins
> Doing this as e-mail aliases is one thing, but doing it for usernames
> just seems like BOFH behavior.
I dunno... it sounds evil enough, but requires enough extra work for
the operator(s) due to aforementioned breakage...
that even Simon himself probably wouldn't stoop this low.
--
Remove the spamhole to reply.
Reinier Post
>I don't know why you're doing this, but it's usually done for the
>wrong reasons. It does not create predictable login names; it makes
>them less unique.
Not only that, it's more typing, and you type your account name a *lot*.
Plus the idea that you can get unique usernames this way eventually
breaks down. It's just tedious nonsense.
>2. Let users pick their login names. They like it better that way
> anyhow. Usually a mandatory naming scheme is resented as a
> kind of bureaucratic anal-retentive tyranny by users - yes, they
> actually do take it personally.
This is because they are reminded every time they read their username,
which is about a thousand times per day.
>Regards,
>
>--Paul
--
Reinier (rp)
I R A Darth Aggie
Alan Coopersmith <al...@CSUA.Berkeley.EDU>, in
<ab79vg$2m4r$1...@agate.berkeley.edu> wrote:
<about long user names, which may well break things>
+> Doing this as e-mail aliases is one thing, but doing it for usernames
+> just seems like BOFH behavior.
Speaking only for myself, as a BOFH, I would never inflict such pain
upon myself. You think I want to have to deal with such nonsense?
The desire for this probably comes down from on high, in the land of
the Pointy-Haired People. Smells like a dead woodchuck.
James
--
Consulting Minister for Consultants, DNRC
I can please only one person per day. Today is not your day. Tomorrow
isn't looking good, either.
I am BOFH. Resistance is futile. Your network will be assimilated.
Charles Soto
Why don't you just assign them 16-digit ISO UINs (Unique
Identification Numbers) specific to your organization? ;-)
Charles
(everyone remembers their Compuserve account numbers, right?)
Ian Gregory
>In my last contract, the user namespace was constrained to 8 characters
>and defined by "Lastname (7 char) + Firstname (1st char)" unless there
>was a collision with a prexisting account, then it was "Lastname (6
>char) + Firstname (1st char) + Middle Initial or Firstname (2nd char)".
>There were some collisions in the Ph* Nuygen group of users and we
>worked it out on first-come, first-name basis.
As I see it the two main requirements for usernames are:
1) Short, easy to generate without uniqueness issues and
easy for user to remember.
2) Difficult for anyone else to deduce the identity of a person from
user name or vica versa - UNLESS they have access to a directory.
If managment wants random sociopaths to be able to send email to any
of your users knowing only their name I would be worried about privacy
issues and cases of mistaken identity.
Anyway, this functionality would be better implemented with a directory
search tool on the company website with a clear privacy policy regarding
exactly what information should be revealed.
I like the username scheme xp2023 jh9263 kk2274 etc
This seems to be a good compramise between 1) and 2), specially
if you make the letters be the user's initials, and the number be
essentially random (people are used to remembering 4 digit PINs).
--
Ian Gregory
Systems and Applications Manager
Learning and Information Services
University of Hertfordshire
Benny
> Identification Numbers) specific to your organization? ;-)
Or tattoo a barcode on their forehead and have them read those with a
barcode scanner (at the entrance of their cubicle) giving them instant
access ?
A tad hard to change the password and clearly they need to wear hats all
the time because someone may take a photograph. But otherwise it's
fullproof ;-P
* wink *
Benny
ps: nice list of newsgroups you post togheter. ( HI WORLD ! )
YTC#1
just keep it simple, 1st and last initial, then a number.
--
Bruce Porter
XJR1300SP, XJ900F, GSX750W, GS550, GSX250, CB175
WUSS#1 , YTC#1(bar), OSOS#2 , DS#3 , IbW#18
"The internet is a huge and diverse community
and not every one is friendly"
http://www.ytc1.co.uk
Jeremiah DeWitt Weiner
> I like the username scheme xp2023 jh9263 kk2274 etc
> This seems to be a good compramise between 1) and 2), specially
> if you make the letters be the user's initials, and the number be
> essentially random (people are used to remembering 4 digit PINs).
Ugh. And people are supposed to remember each others' numbers too?
That 'js6326' is Joe Smith in accounting, not John Smith in IT? Or vice
versa, that Kelly Cline in Operations is 'kc7423', not 'kc8947'? (Then
again, users somehow manage to remember 50 random ten-digit phone numbers
effortlessly, but can't remember their one single password, which they
chose themselves, so sometimes generalizing about memory seems hopeless.)
I mean, if you're going to have that much entropy in your logins it seems
to me like you might as well bite the bullet and make them completely
numeric.
My feeling is that one should have a policy, but a very flexible
one. (I recall the old story about the place that decided to use
first initial + first seven letters of last name. It was pointed out
to them that Steve Hittinger would not like this. They changed it to
first seven letters of last name + first initial. It was pointed out
that Ken Chin would not like this. They threw up their hands in despair.)
First initial + middle initial + first six of last name works pretty well.
If you have namespace collisions, use first six of first name + last
initial. It isn't a matter of life and death, you know. (My current
company uses 5+3, eg "George Bush" gets "georgbus"! Guess who invented
/that/ scheme...)
--
If mail to me bounces, try removing the "+STRING" part of the address.
Lance
news:ab9bfh$1ik$1...@reader1.panix.com...
> Ugh. And people are supposed to remember each others' numbers too?
> That 'js6326' is Joe Smith in accounting, not John Smith in IT? Or vice
> versa, that Kelly Cline in Operations is 'kc7423', not 'kc8947'? (Then
> again, users somehow manage to remember 50 random ten-digit phone numbers
> effortlessly, but can't remember their one single password, which they
> chose themselves, so sometimes generalizing about memory seems hopeless.)
> I mean, if you're going to have that much entropy in your logins it seems
> to me like you might as well bite the bullet and make them completely
> numeric.
You can have aliases for email. Where Joe.S...@company.com goes to js6326
Lance
Paul Southworth
Vomit. And Joe.Sm...@company.com goes to js4040 and
Joe.Sm...@company.com goes to js2121 - OK which is which? I
forget - was Joe.Smitch.4 the one whose daughter died in a car
wreck, and Joe.Smitch.2 just had a daughter who had a baby? Oh
well, copy them all and you'll know the message got through.
Who is number one?
Why do people have such a hard time just letting users pick their
own login names? It easy, the staff will be happy, you won't run
out of login names, and others in the company will actually remember
their coworkers email addresses, often using them to the exclusion of
given names.
--Paul
those who know me have no need of my name
>(everyone remembers their Compuserve account numbers, right?)
yes.
--
bringing you boring signatures for 17 years
Lance
news:ol_B8.19140$A%3.18...@ord-read.news.verio.net...
> >You can have aliases for email. Where Joe.S...@company.com goes to
js6326
>
> Vomit. And Joe.Sm...@company.com goes to js4040 and
> Joe.Sm...@company.com goes to js2121 - OK which is which? I
> forget - was Joe.Smitch.4 the one whose daughter died in a car
> wreck, and Joe.Smitch.2 just had a daughter who had a baby? Oh
> well, copy them all and you'll know the message got through.
> Who is number one?
>
> Why do people have such a hard time just letting users pick their
> own login names? It easy, the staff will be happy, you won't run
> out of login names, and others in the company will actually remember
> their coworkers email addresses, often using them to the exclusion of
> given names.
>
> --Paul
There is a difference between letting them pick their login names and
letting them pick their email names or having email names with a lot of
numbers.
My point was to the previous poster that mentioned number in the email
address. Someone in the company wants to send Joe Smith an email. Do you
think they are going to know that his email is jos5...@company.com If Joe
Smith wanted jos58358 as his login name, that fine. On the email server you
can have an alias of Joe....@company.com to go to jos58358 Then Joe is
happy and people that need to email him are happy. Having an email address
that is easy to remember can be very helpful, especially if they are in
sales. After all, if Joes' customers keep getting a message from the
postmaster that their message could not be delivered, they may go elsewhere.
Every company I have worked for, my login name was not the same an my email
name.
Lance
Steve Bellenot
Paul Southworth <cnhyf-10...@usenet.etext.org> wrote:
>
>Why do people have such a hard time just letting users pick their
>own login names? It easy, the staff will be happy, you won't run
>out of login names, and others in the company will actually remember
>their coworkers email addresses, often using them to the exclusion of
>given names.
>
Maybe it has something to do with the number of users or the timing.
What works for a handful of people of friendly folks becomes unworkable
fairly fast as the numbers grow or as the distances grow. Some colleges
generate accounts for potential freshmen as soon as they get the first
piece of paper (the one with the check no doubt) in the application.
You get an account before you get accepted.
Many (most?) people don't seem to type email addresses. They reply to
old messages. Or what amounts to the same, use an address book formed
from old messages.
--
http://www.math.fsu.edu/~bellenot
bellenot <At/> math.fsu.edu
+1.850.644.7189 (4053fax)
Lance
news:aba69g$614$1...@news.fsu.edu...
> Maybe it has something to do with the number of users or the timing.
> What works for a handful of people of friendly folks becomes unworkable
> fairly fast as the numbers grow or as the distances grow. Some colleges
> generate accounts for potential freshmen as soon as they get the first
> piece of paper (the one with the check no doubt) in the application.
> You get an account before you get accepted.
>
> Many (most?) people don't seem to type email addresses. They reply to
> old messages. Or what amounts to the same, use an address book formed
> from old messages.
Yes, but if a person is a sales person, sometimes and easy email address is
the best. Also, what one person logins with and what their email address
can cause a big problem. Here are two examples. The first name is John
Long. He decides his login name is longjohn The company could have sexual
harassment suits because of that. Or he is another one. This guy decides
just to use his last name as his login ID: FOCKER
That's fine if that is what he wants to login with, email is a different
story.
Lance
Mark Campbell
news:slrnadg3dq.i9...@ihg0.herts.ac.uk...
> Michael Vilain wrote:
>
> >In my last contract, the user namespace was constrained to 8 characters
> >and defined by "Lastname (7 char) + Firstname (1st char)" unless there
> >was a collision with a prexisting account, then it was "Lastname (6
> >char) + Firstname (1st char) + Middle Initial or Firstname (2nd char)".
> >There were some collisions in the Ph* Nuygen group of users and we
> >worked it out on first-come, first-name basis.
>
Sorry I'm jumping in mid thread here, so sorry if I'm making a point some
one has allready made :)
The main problem I could envisage with a username with a '.' charater in it
is the fact that
alot of shell based programes don't like them in username, write is an
example of this I believe..
The idea of having username being constructed of in the for of mcampbel,
would be wise if you have an
mid sized number of users (under 10 K). This way it reamins personal. In
case of duplicates just add
another letter of the first name and remove the last of the second. You
should use first.second@hotsname for
e-mail addresses, however you should maintain this via /etc/aliases file or
some such.
hope this helped a little :)
-mark
Kjetil Torgrim Homme
>
> "Paul Southworth" <cnhyf-10...@usenet.etext.org> wrote:
> > Vomit. And Joe.Sm...@company.com goes to js4040 and
> > Joe.Sm...@company.com goes to js2121 - OK which is which? I
> > forget - was Joe.Smitch.4 the one whose daughter died in a car
> > wreck, and Joe.Smitch.2 just had a daughter who had a baby? Oh
> > well, copy them all and you'll know the message got through.
> > Who is number one?
>
> and letting them pick their email names or having email names with
> a lot of numbers.
>
> My point was to the previous poster that mentioned number in the
> email address. Someone in the company wants to send Joe Smith an
> email. Do you think they are going to know that his email is
> jos5...@company.com If Joe Smith wanted jos58358 as his login
> name, that fine. On the email server you can have an alias of
> Joe....@company.com to go to jos58358 Then Joe is happy and
> people that need to email him are happy.
you didn't address the problem. what do you do when another Joe Smith
shows up?
Kjetil Torgrim Homme
>
> The idea of having username being constructed of in the for of
> mcampbel, would be wise if you have an mid sized number of users
> (under 10 K). This way it reamins personal. In case of
> duplicates just add another letter of the first name and remove
> the last of the second.
btw, I have a Perl script for this available at
http://www.ifi.uio.no/~kjetilho/hacks/
it's the algorithm we're using at the University of Oslo.
dave...@spamcop.net
> you didn't address the problem. what do you do when another Joe Smith
> shows up?
Ah, that's the easy part. After a week, hold a vote of the users,
and decide which one to keep, and which to fire.
There's no problem that can't be solved, if you're creative enough.
Dave
I R A Darth Aggie
Kjetil Torgrim Homme <kjet...@haey.ifi.uio.no>, in
<1roffqa...@fimm.ifi.uio.no> wrote:
+> [Lance]:
+> >
+> > "Paul Southworth" <cnhyf-10...@usenet.etext.org> wrote:
+> > > Vomit. And Joe.Sm...@company.com goes to js4040 and
+> > > Joe.Sm...@company.com goes to js2121 - OK which is which? I
+> > > forget - was Joe.Smitch.4 the one whose daughter died in a car
+> > > wreck, and Joe.Smitch.2 just had a daughter who had a baby? Oh
+> > > well, copy them all and you'll know the message got through.
+> > > Who is number one?
+> >
+> > There is a difference between letting them pick their login names
+> > and letting them pick their email names or having email names with
+> > a lot of numbers.
+> >
+> > My point was to the previous poster that mentioned number in the
+> > email address. Someone in the company wants to send Joe Smith an
+> > email. Do you think they are going to know that his email is
+> > jos5...@company.com If Joe Smith wanted jos58358 as his login
+> > name, that fine. On the email server you can have an alias of
+> > Joe....@company.com to go to jos58358 Then Joe is happy and
+> > people that need to email him are happy.
+>
+> you didn't address the problem. what do you do when another Joe Smith
+> shows up?
Joe.Smith.the.dumb.ass.marketing.dude@f*ckedcompnay.com
Ian Gregory
>In comp.sys.sun.admin Ian Gregory <I.H.G...@herts.ac.uk> wrote:
>> I like the username scheme xp2023 jh9263 kk2274 etc
>
>> This seems to be a good compramise between 1) and 2), specially
>> if you make the letters be the user's initials, and the number be
>> essentially random (people are used to remembering 4 digit PINs).
>
> Ugh. And people are supposed to remember each others' numbers too?
NO NO NO - that is my whole point! You never need to REMEMBER anyone
else's username, leave the remembering to the machines.
I am talking about systems with a lot of users. I want to email
my friend Simon whose last name happens to be Patel. Now is he
"S.37.Patel" or "S.39.Patel"? Oh dear I can't remember. Doesn't matter
I just send an email to "Simon" because he is in my address book.
If I don't already have his username then do I try emailing all
40 users called S[.][0-9]*[.]Patel ?
No, with a sensible system I look up his name in a directory. If there
are multiple hits I refine the search (eg I know he is in Engineering).
Assuming I succeed in finding a single match and an email address is
listed I just put it in my address book under "Simon".
It makes absolutely no difference to me whether it goes in as
"S.39.Patel" or "sp1764" because to me he is just "Simon".
Imagine trying to get rid of phone numbers because they are hard to
remember. Why not just use people's names instead? Well, how do you
distinguish between the 10,000 John Smiths who have no other initial.
Call them John-Smith-0000, John-Smith-0001 etc? The whole idea is
ludicrous.
There are clowns who don't understand this and set up useless email
aliases and then there are the idiots who try to use them for UNIX
usernames. A pox on them all!
Jerry Leslie
:
: you didn't address the problem. what do you do when another Joe Smith
: shows up?
:
I faced that problem at a previous job where I was the email admin
for the local linux/POP3 mail server.
Aliases were used to provide the corporate standard:
<firstname>.<lastname>
When we had a conflict then we used:
<firstname>.<middleinitial><lastname>
The first corporate standard email system was Oracle Office, which our
site refused to use because it was an ill-behaved 16-bit app that could
bring down Windows NT. Then Oracle Office was replaced with M$ Exchange.
--Jerry Leslie les...@clio.rice.edu (my opinions are strictly my own)
Note: les...@jrlvax.houston.rr.com is invalid for email
Lance
news:1roffqa...@fimm.ifi.uio.no...
> > My point was to the previous poster that mentioned number in the
> > email address. Someone in the company wants to send Joe Smith an
> > email. Do you think they are going to know that his email is
> > jos5...@company.com If Joe Smith wanted jos58358 as his login
> > name, that fine. On the email server you can have an alias of
> > Joe....@company.com to go to jos58358 Then Joe is happy and
> > people that need to email him are happy.
>
> you didn't address the problem. what do you do when another Joe Smith
> shows up?
Let the users assign their own login name, have corporate assign the email
address. What would happen if the first Joe Smith used Joe.Smith and the
second one wanted the same thing.
Lance
Will Packard
>
<snip>
> The first corporate standard email system was Oracle Office, which our
> site refused to use because it was an ill-behaved 16-bit app that could
> bring down Windows NT. Then Oracle Office was replaced with M$ Exchange.
<sarcasm>
Oh, that's *much* better.
</sarcasm>
Will
Paul Southworth
Steve Bellenot <bell...@math.fsu.edu> wrote:
>In article <ol_B8.19140$A%3.18...@ord-read.news.verio.net>,
>Paul Southworth <cnhyf-10...@usenet.etext.org> wrote:
>>
>>Why do people have such a hard time just letting users pick their
>>own login names? It easy, the staff will be happy, you won't run
>>out of login names, and others in the company will actually remember
>>their coworkers email addresses, often using them to the exclusion of
>>given names.
>>
>
>Maybe it has something to do with the number of users or the timing.
>What works for a handful of people of friendly folks becomes unworkable
>fairly fast as the numbers grow or as the distances grow. Some colleges
>generate accounts for potential freshmen as soon as they get the first
>piece of paper (the one with the check no doubt) in the application.
There are also really big sites that let people select it, too,
universities and of course ISPs as well.
How long does it take to write a web app that lets people type in
names, checks if they are valid/existant, and proposes some
alternatives? That can be implemented in an hour by even an
incompetent web programmer. It just isn't any extra work.
>You get an account before you get accepted.
>
>Many (most?) people don't seem to type email addresses. They reply to
>old messages. Or what amounts to the same, use an address book formed
>from old messages.
Or an institutional directory...
--Paul
p...@icke-reklam.ipsec.nu
> "Paul Southworth" <cnhyf-10...@usenet.etext.org> wrote in message
> news:ol_B8.19140$A%3.18...@ord-read.news.verio.net...
>> >You can have aliases for email. Where Joe.S...@company.com goes to
> js6326
>>
>> Vomit. And Joe.Sm...@company.com goes to js4040 and
>> Joe.Sm...@company.com goes to js2121 - OK which is which? I
>> forget - was Joe.Smitch.4 the one whose daughter died in a car
>> wreck, and Joe.Smitch.2 just had a daughter who had a baby? Oh
>> well, copy them all and you'll know the message got through.
>> Who is number one?
>>
>> Why do people have such a hard time just letting users pick their
>> own login names? It easy, the staff will be happy, you won't run
>> out of login names, and others in the company will actually remember
>> their coworkers email addresses, often using them to the exclusion of
>> given names.
>>
>> --Paul
> There is a difference between letting them pick their login names and
> letting them pick their email names or having email names with a lot of
> numbers.
> My point was to the previous poster that mentioned number in the email
> address. Someone in the company wants to send Joe Smith an email. Do you
You just don't guess an e-mail name. You get it from some source,
could be "sender or reply-to address", it can be something printed
on a buisniess card, or it could be a direcyory on a webserver.
Get a better argument why people cannot state their usernames !
> think they are going to know that his email is jos5...@company.com If Joe
> Smith wanted jos58358 as his login name, that fine. On the email server you
> can have an alias of Joe....@company.com to go to jos58358 Then Joe is
> happy and people that need to email him are happy. Having an email address
> that is easy to remember can be very helpful, especially if they are in
> sales. After all, if Joes' customers keep getting a message from the
> postmaster that their message could not be delivered, they may go elsewhere.
> Every company I have worked for, my login name was not the same an my email
> name.
You have worked on the wrong companies :-)
> Lance
--
Peter Håkanson
IPSec Sverige ( At Gothenburg Riverside )
Sorry about my e-mail address, but i'm trying to keep spam out,
remove "icke-reklam" if you feel for mailing me. Thanx.
Jeremiah DeWitt Weiner
> Aliases were used to provide the corporate standard:
> <firstname>.<lastname>
> When we had a conflict then we used:
> <firstname>.<middleinitial><lastname>
So what if you have two Joe R. Smiths? What if you don't know Joe
Smith in Marketing's middle initial? Or what if you do, but you know
there's more than one Joe Smith - how do you know if he's the one who got
the middle initial or not?
Jeremiah DeWitt Weiner
> No, with a sensible system I look up his name in a directory. If there
> are multiple hits I refine the search (eg I know he is in Engineering).
> Assuming I succeed in finding a single match and an email address is
> listed I just put it in my address book under "Simon".
So with the one system, you just know, and with the other, you're
given a cryptic string that you have to search for to get any meaning
out of it - assuming your organization is even very good at keeping its
directory up to date, which most are not. I'm not sold. Look, I agree
that in very large organizations, username namespace does become a problem.
But it becomes a problem no matter what you do, and you can try to solve it
in a human-friendly way, or a human-unfriendly way.
(As for whether we should get rid of phone numbers, a lot of
usability experts think we should! Come on - if anybody invented a new
system, where each user was identified only by a semirandom combination of
digits anywhere from 4 to 14 digits long, which had only very limited and
non-centralized one-way search capability, you'd tell them to go shove it.)
Lack Mr G M
|>
|> So with the one system, you just know,
No, you *might* know. What you might know is another question...
|> and with the other, you're
|> given a cryptic string that you have to search for to get any meaning
|> out of it
No you do not. You send mail to it, if you wish. You will be given
it in a context that describes it, eg: on a business card, in an SMTP
header along with a display name.
|> - assuming your organization is even very good at keeping its
|> directory up to date, which most are not.
So fix the problem...
|> But it becomes a problem no matter what you do, and you can try to solve it
|> in a human-friendly way, or a human-unfriendly way.
Having multiple John.Smith entries is not friendly. Or is it
Jack.Smith? All very confusing and not friendly. And I might not want
you to be able to mail me just because you know my name.
|> where each user was identified only by a semirandom combination of
|> digits anywhere from 4 to 14 digits long, which had only very limited and
|> non-centralized one-way search capability, you'd tell them to go shove it.)
No, I'd tell you to have a centralized (or rather globally available)
two-way search capability.
|> If mail to me bounces, try removing the "+STRING" part of the address.
And there was me thinking that your name was "jdw ALLSPAMMERSMUSTDIE".
--
--------- Gordon Lack --------------- gml...@ggr.co.uk ------------
This message *may* reflect my personal opinion. It is *not* intended
to reflect those of my employer, or anyone else.
Lance
news:abdov3$54h$4...@nyheter.crt.se...
> You just don't guess an e-mail name. You get it from some source,
> could be "sender or reply-to address", it can be something printed
> on a buisniess card, or it could be a direcyory on a webserver.
I guess no one seems to TALK anymore. There's a way to get an email
address, people write tings down wrong. If you start throwing numbers in
it, good luck. Or worse yet, two letters and then numbers. At which point,
if someone just wrote down the email address. they may have no clue who it
next week.
>
> Get a better argument why people cannot state their usernames !
Usernames do not have to be the same as email addresses. Quit confusing the
two.
> > Every company I have worked for, my login name was not the same an my
email
> > name.
>
> You have worked on the wrong companies :-)
Your email was first...@compnay.com Easy for other people to remember,
but the login was different. Very few companies have their entire network
all tied in. Various systems do not use authentication as others.
Lance
p...@icke-reklam.ipsec.nu
> <p...@icke-reklam.ipsec.nu> wrote in message
> news:abdov3$54h$4...@nyheter.crt.se...
>> You just don't guess an e-mail name. You get it from some source,
>> could be "sender or reply-to address", it can be something printed
>> on a buisniess card, or it could be a direcyory on a webserver.
> I guess no one seems to TALK anymore. There's a way to get an email
> address, people write tings down wrong. If you start throwing numbers in
> it, good luck. Or worse yet, two letters and then numbers. At which point,
> if someone just wrote down the email address. they may have no clue who it
> next week.
>>
>> Get a better argument why people cannot state their usernames !
> Usernames do not have to be the same as email addresses. Quit confusing the
> two.
I'm not confusing this. However discussion was about unix account names.
Thats what i was talking about.
> > > Every company I have worked for, my login name was not the same an my
> email
>> > name.
>>
>> You have worked on the wrong companies :-)
> Your email was first...@compnay.com Easy for other people to remember,
> but the login was different. Very few companies have their entire network
> all tied in. Various systems do not use authentication as others.
Still no reason why a new user cannot suggest a userid for himself.
As long as it passes (whatever rules you might see) i cannot see
any reason why a person that happens to be known by nickname "jack" cannot
suggest "jack" as userid. If it happens to be occupied, then find
out something else.
Lance
news:abejhj$ogj$2...@nyheter.crt.se...
> I'm not confusing this. However discussion was about unix account names.
> Thats what i was talking about.
Then the thread turned to email addresses as well.
>
> > > > Every company I have worked for, my login name was not the same an
my
> > email
> >> > name.
> >>
> >> You have worked on the wrong companies :-)
>
> Still no reason why a new user cannot suggest a userid for himself.
> As long as it passes (whatever rules you might see) i cannot see
> any reason why a person that happens to be known by nickname "jack" cannot
> suggest "jack" as userid. If it happens to be occupied, then find
> out something else.
They should be able to use what ever name they want. there are people with
the name Gerald that use Jerry. Even a Tom that like Bob.
But using a complicated to others user id as an email address is not very
user friendly.
Lance
Damian Ogorow
> cnhyf-10...@usenet.etext.org (Paul Southworth) wrote:
>
>
>>I don't know why you're doing this, but it's usually done for the
>>wrong reasons. It does not create predictable login names; it makes
>>them less unique.
>
ITS SIMPLE
/etc/aliases - for e-mail adres
and
/etc/passwd - what You like in 8 char.
Damian
srl...@read.the.sig
Kjetil Torgrim Homme <kjet...@haey.ifi.uio.no> writes:
>you didn't address the problem. what do you do when another Joe Smith
>shows up?
Then the Joe....@company.com email alias becomes JoeM.Smith and the new
Joe Smith becomes JoeD.Smith (where M and D are middle initials for example).
After a short period, mail to Joe....@company.com is bounced (ideally
with a helpful message pointing out the two choices).
This way "easy-to-remember" mail addresses can continue to be used for the
other 999 people in the company. There's a minor (and short term) inconvenience
for the first Joe Smith while he gets his business card reprinted and tells
his correspondents that he has a new mail address.
We do this and have had only 2 or 3 clashes out of 400 or so people. This
is with using just a J.S...@company.com style too.
--
Roger Williams, Institute of Geological & Nuclear Sciences, New Zealand
R.Williams
@gns.cri.nz : DEC 3000/300 AXP - OpenVMS v6.2
Jim Patterson
news:michael-D428F2...@news.tdl.com...
> In article <abejhj$ogj$2...@nyheter.crt.se>, p...@icke-reklam.ipsec.nu
>
> The last company I worked at had over 3000 PCs scattered throught the
> company. People were given a PC + NT account + a NOVELL fileshare and
> access to the various corporate UNIX systems as needed. All these
> systems used the same username namespace, so it had to match across all
> of them. We had some people who wanted Jerry rather than Gerrald which
> was on their badge. The badge won, end of discussion. IS didn't have
> the people to manage all the disparte names unless they all conformed to
> a standard.
>
> Corporate email was another matter. Entries were placed in aliases to
> map entries like last....@company.com, user...@company.com,
> first...@company.com, and any special cases. Email had 2 full-time
> people managing it (along with calendaring). There were still M$
> Exchange systems and W95 systems on the net in various places that
> required gateways for these groups.
>
> If you're at a company that lets you select your own username and host
> name, great. Lots of big companies don't have the smarts or
> infrastructure or people resources to do that.
>
> The orginal poster's request was how to do first.last for UNIX
> usernames. Some Unix systems will allow long usernames and passwords.
> Solaris will 'sort of' allow it, but it's not Best Practice. Stay with
> the 8 character usernames for that namespace if you're using UNIX
> systems.
>
> --
> Michael Vilain Certified Advanced Rolfer
> rol...@vilain.com http://www.vilain.com
>
> Envision a just and peaceful world, in which all people achieve their full
potential and live with love, respect, and compassion.
At the DOD plant where I work for a subcontractor with a total of about 3000
people on base, most of our computer systems (including the UNIX systems)
use our employee ID:single letter plus 5-digit integer to logon. Our email
is first.last name, where first name may be actual first name or
nickname, whichever you prefer.
--
Jim Patterson
Kjetil Torgrim Homme
>
> Kjetil Torgrim Homme <kjet...@haey.ifi.uio.no> writes:
> > you didn't address the problem. what do you do when another Joe
> > Smith shows up?
>
> Then the Joe....@company.com email alias becomes JoeM.Smith and
> the new Joe Smith becomes JoeD.Smith (where M and D are middle
> initials for example).
whoa, back up! why do you assume they have initials? I didn't
mention an initial. we have four Thomas Hansen at the Department of
Informatics. no middle name. (we have another Thomas Hansen _with_
middle names.)
> After a short period, mail to Joe....@company.com is bounced
> (ideally with a helpful message pointing out the two choices).
>
> This way "easy-to-remember" mail addresses can continue to be used
> for the other 999 people in the company. There's a minor (and
> short term) inconvenience for the first Joe Smith while he gets
> his business card reprinted and tells his correspondents that he
> has a new mail address.
an inconvenience which is totally unnecessary.
> We do this and have had only 2 or 3 clashes out of 400 or so
> people. This is with using just a J.S...@company.com style too.
good for you.
Toni Arte
> > Smith in Marketing's middle initial? Or what if you do, but you know
> > there's more than one Joe Smith - how do you know if he's the one who got
> > the middle initial or not?
>
>
> Joe.Smith
> Joe.RSmith
> J.RSmith
> J.Smith
> Smith.JR
> Smith.JoeR
> Smith.j
>
> If you have more than seven people called Joe R Smith, then I suggest
> <user>@marketing.company.com
> <user>@production.company.com
In our site we have three people with exactly the same first, middle and
last name :-) So the email addresses are like this:
* the first one has the first letter from the middle name
* the second one has two letters from the middle name
* the third one has three letters from the middle name
Like this:
Joe.R.Smith
Joe.Ri.Smith
Joe.Ric.Smith
--
Toni
p...@icke-reklam.ipsec.nu
> Jeremiah DeWitt Weiner <jdw+ALLSPAM...@panix.com> writes:
>> So what if you have two Joe R. Smiths? What if you don't know Joe
>> Smith in Marketing's middle initial? Or what if you do, but you know
>> there's more than one Joe Smith - how do you know if he's the one who got
>> the middle initial or not?
> Oh for heaven's sake.
> Joe.Smith
> Joe.RSmith
> J.RSmith
> J.Smith
> Smith.JR
> Smith.JoeR
> Smith.j
> If you have more than seven people called Joe R Smith, then I suggest
> <user>@marketing.company.com
> <user>@production.company.com
As e-mail identitys :
don't forget that the best way of enshuring that all these
"john.smiths" are satisfied with their acronymes is to
actually let them choose.
The discussion was about userid, where capitalization and dot's are
often prohibited. Again, letting the johns decide and suggest what
they are comfortable with a the best way.
So instead of letting clueless admins and corporate "standards" dictate
userid, find ways to give users influence about this.
Many people are picky about what they are called. Maybe not all
can go so far as i once did, stating "the left side of my emailaddress
contains "peter", the right is up to you". Of course i got it my way,
proudly keepeing peter on my left side ( unless there was a previous
occumant, in those cases i choose another acronym).
One's name is importent !
> or something.
> J
> --
> jmd <j...@nelefa.org>
Geoff Lane
> Smith wanted jos58358 as his login name, that fine. On the email server you
> can have an alias of Joe....@company.com to go to jos58358 Then Joe is
> happy and people that need to email him are happy. Having an email address
> that is easy to remember can be very helpful, especially if they are in
> sales. After all, if Joes' customers keep getting a message from the
> postmaster that their message could not be delivered, they may go elsewhere.
What happens when Joe Smith leaves and a different Joe Smith joins the
company? Unfortunately Joe Smith (1) was signed up with a number of
pornographic joke email lists (the reason he "left" the company in the first
place) that won't or can't delete his address. Joe Smith (2) gets all that
garbage plus whinny emails from Joe Smith (1)'s ex-girfriend who will not
believe he has left and keeps adding his name to pornographic joke email
lists or worse because he won't call her.
Real world names in email addresses are not a good idea.
--
/\ Geoff. Lane. /\ Manchester Computing /\ Manchester /\ M13 9PL /\ England /\
I've used Basic so long, my brain has gonesub permanently
Mike Marshall
>Joe.R.Smith
>Joe.Ri.Smith
>Joe.Ric.Smith
I don't know what his middle name is, but I want to send email
to one of them. I guess I'll just send it to all three. No, wait
a minute, it is personal, I don't want anyone else to see it, I'll
pick Joe.Ric.Smith - D'OH! wrong one. Turns out Joe.R.Smith is
tired of getting Joe.Ri.Smith's email, so he's just started
deleting it.
http://hubcap.clemson.edu/~hubcap/d.weekly/firstname.lastname.html
-Mike
Ian Gregory
>
>My name is John. If I worked at a company where there were four johns and
>someone said "sorry, dude, we don't want to get confused so we're going
>to call you jack" I'd be more than a little miffed.
But if they gave you a telephone and told you your extension number
was "3033" would you be miffed? Would you demand that they change
your phone number to "John"?
dave...@spamcop.net
> I don't know what his middle name is, but I want to send email
> to one of them. I guess I'll just send it to all three. No, wait
> a minute, it is personal, I don't want anyone else to see it, I'll
> pick Joe.Ric.Smith - D'OH! wrong one. Turns out Joe.R.Smith is
> tired of getting Joe.Ri.Smith's email, so he's just started
> deleting it.
Bah. It's a one-time learning curve issue. Sure, the mail will get
to the wrong guy once in a while, but that'd happen no matter what
scheme you use.
"Hey Joe, which one of these guys is you? Oh, OK, thanks." Not
a hard thing to resolve for the users. Beats the hell out of having
to remember who nicdjh0 or whatever is.
those who know me have no need of my name
>[srl...@read.the.sig]:
>> This way "easy-to-remember" mail addresses can continue to be used
>> for the other 999 people in the company. There's a minor (and
>> short term) inconvenience for the first Joe Smith while he gets
>> his business card reprinted and tells his correspondents that he
>> has a new mail address.
>
>an inconvenience which is totally unnecessary.
and can be expensive, depending on the person's position and the turn-over
rate. some people only have a limited number of business cards, so it
might be cheap (that one time), but what if the original person not only
has lots of expensive cards but also some stationary (e.g., is an exec).
--
bringing you boring signatures for 17 years
Reinier Post
>The last company I worked at had over 3000 PCs scattered throught the
>company. People were given a PC + NT account + a NOVELL fileshare and
>access to the various corporate UNIX systems as needed. All these
>systems used the same username namespace, so it had to match across all
>of them. We had some people who wanted Jerry rather than Gerrald which
>was on their badge. The badge won, end of discussion. IS didn't have
>the people to manage all the disparte names unless they all conformed to
>a standard.
I don't understand this argument. Usernames cannot be predicted from
the user's real name. You can devise a scheme that works for 99% of
cases, but you'll still need to pick the rest by hand. So there must
be a hand-made table for these 1% of users. It can't be much more work
to maintain the table for all 100% of users.
A more relevant difference is the fact that if you let users pick
their own account names, you must actually do this; you can only
create the account once the user has picked the name. However, you
probably want the user's explicit confirmation first before creating
the account anyway, so asking for the preferred name isn't going to be
much extra work.
--
Reinier
Ian Gregory
So how did you talk to Joe? By phone? Where did you get his phone
number? From a directory?
So the directory gives his phone number as whatever and his email
as nicdjh0, or as "not available" if Joe doesn't want random
people emailing him, or it tells you that he has recently left
the company, or let's say he has changed his name to Joleen but
that doesn't matter because her email address is still nicdjh0
and she doesn't need to inform anyone about a change.
Once you have got her email address you file it under "Joleen"
Beats the hell out of emailing Joe.Ri...@crappycorp.com only
to have it bounce.
Anyway if Joe is really THAT bothered he can always forward
j...@joe.com to his work address.
dave...@spamcop.net
> dave...@spamcop.net wrote:
>>
>>"Hey Joe, which one of these guys is you? Oh, OK, thanks." Not
>>a hard thing to resolve for the users. Beats the hell out of having
>>to remember who nicdjh0 or whatever is.
> So how did you talk to Joe? By phone? Where did you get his phone
> number? From a directory?
Yeah, a directory which gives his name, location, phone, department info,
email address, that sort of thing. Trivial to setup and distribute.
> So the directory gives his phone number as whatever and his email
> as nicdjh0, or as "not available" if Joe doesn't want random
> people emailing him, or it tells you that he has recently left
> the company, or let's say he has changed his name to Joleen but
> that doesn't matter because her email address is still nicdjh0
> and she doesn't need to inform anyone about a change.
Well, we have people whose last names change regularly; if they want
it changed, we'll change it. uid stays the same, login changes,
put an alias in /etc/aliases pointing smithj to jonesj, and we're
done.
> Once you have got her email address you file it under "Joleen"
> Beats the hell out of emailing Joe.Ri...@crappycorp.com only
> to have it bounce.
Whatever. How often do your users have sex-change operations that
this is really a burning issue?
> Anyway if Joe is really THAT bothered he can always forward
> j...@joe.com to his work address.
Introducing an external dependancy to a business email path? Not
on my network; my directory will certainly not give an external address
for someone who has an internal one.
How we've handled it is this, in order of collisions:
lastname
lastnamefirstinitial
lastnamefirstinitialsecondinitial
"Hey John L. Schmidt, we've already got a schmidtjl, what would you like
your login and address to be?". However, I stand by my earlier statement
that often, it's easier just to hold a vote and decide which person to
keep, and get rid of one of 'em.
Dave Hinz
Mike Marshall
>"Hey Joe, which one of these guys is you? Oh, OK, thanks."
How did you ask him that, call him? Is he in the same office as
you? Who needs email anyhow? <g>
We use that thingamabob in sendmail that hooks into qi so that when you
send email to firstname.lastname@here you get the results
of a ph query sent back to you. That's how you find out that
joe.smith in the ag department = xj1179@here...
-Mike
Kai Großjohann
> In comp.sys.sun.admin Ian Gregory <I.H.G...@herts.ac.uk> wrote:
>> No, with a sensible system I look up his name in a directory. If there
>> are multiple hits I refine the search (eg I know he is in Engineering).
>> Assuming I succeed in finding a single match and an email address is
>> listed I just put it in my address book under "Simon".
>
> So with the one system, you just know, and with the other, you're
> given a cryptic string that you have to search for to get any meaning
> out of it - assuming your organization is even very good at keeping its
> directory up to date, which most are not.
I thought the directory would be the authoritative source on login
names, too? Ie, if the directory is LDAP, then let the login process
use LDAP, too. If the directory is NIS, then, well, err. It becomes
a bit more inconvenient, but there _is_ the GECOS field...
kai
--
Silence is foo!
Jerry Leslie
: Will Packard <wi...@nortelnetworks.com> writes:
:
: > Jerry Leslie wrote:
:
: [...]
:
: > > The first corporate standard email system was Oracle Office, which our
: > > site refused to use because it was an ill-behaved 16-bit app that could
: > > bring down Windows NT. Then Oracle Office was replaced with M$ Exchange.
: > <sarcasm>
: > Oh, that's *much* better.
: > </sarcasm>
: It is. Sexchange *is* an ill-behaved 32-bit app, so that's at least an
: improvement ;)
:
Oracle abandoned Oracle InterOffice:
http://news.com.com/2100-1001-221719.html
Oracle switches out own email system - Tech News - CNET.com
By Erich Luening
Staff Writer, CNET News.com
February 17, 1999, 2:30 PM PT
"update In dumping its own internal client/server email system and
switching to an Internet-based messaging platform, Oracle is simply
following its own market advice.
In a company-wide memo to employees, Oracle chief executive Larry
Ellison announced that the company was ending its use of Oracle
InterOffice client/server messaging software and moving to Web-based
email with Netscape Communications' Communicator, a company
spokesperson confirmed.
The Netscape Communicator email client will sit upon Oracle's own
Internet Messaging product, which replaced InterOffice as Oracle's
messaging product last year..."
But it's still in use at Houston Community College:
http://hccmail.hccs.cc.tx.us/
Information Technology Portal
"Monthly IT Availability Report
Network Operational
Internet Operational
Telephone Operational
Oracle E-mail Operational <<<====
IMAP Operational
SPIN Update 05/10/02
PeopleSoft Finance Operational
PeopleSoft HRMS Operational"
http://hccmail.hccs.cc.tx.us/announce/delete.html
EMail Deletion Policy Announcement
"To ensure our Oracle InterOffice/Internet Messaging email system
functions as efficiently as possible for all of our customers, the
following will be implemented beginning Friday, June 23,
2000:
Policy: All UNREAD email located in your INBOX folder older than 60
days will automatically be deleted from the system.
If you do not access your INBOX in more than 120 days, your account
will be deleted. You will need to contact the Information Center to
request a new account..."
--Jerry Leslie les...@clio.rice.edu (my opinions are strictly my own)
Note: les...@jrlvax.houston.rr.com is invalid for email
Toni Arte
>
> Toni Arte <Toni...@nospam.nokia.com> writes:
> >Joe.R.Smith
> >Joe.Ri.Smith
> >Joe.Ric.Smith
>
> I don't know what his middle name is, but I want to send email
> to one of them. I guess I'll just send it to all three. No, wait
> a minute, it is personal, I don't want anyone else to see it, I'll
> pick Joe.Ric.Smith - D'OH! wrong one. Turns out Joe.R.Smith is
> tired of getting Joe.Ri.Smith's email, so he's just started
> deleting it.
Yes, they all say it is a problem. Occasionally I have to email one of
them, and still after a couple of years I have trouble remembering which
'Joe Smith' I have to select as the recipient.
--
Toni
j...@dowcarter.com
> In article <abglfi$14nq$3...@dunlop.dowcarter.com>, John Dow wrote:
>>
>>My name is John. If I worked at a company where there were four johns and
>>someone said "sorry, dude, we don't want to get confused so we're going
>>to call you jack" I'd be more than a little miffed.
> But if they gave you a telephone and told you your extension number
> was "3033" would you be miffed? Would you demand that they change
> your phone number to "John"?
Eh? Telephone numbers are traditionally, well numbers. I wouldn't
expect a telephone number to be "John". I wouldn't expect it to be
a nice shade of lilac. I'd expect it to be a number.
Email addresses tend to be text. Hence "email address" and not
"email number".
I'm not sure your analogy works terribly well ;-)
J
--
John M Dow <j...@dowcarter.com>
Systems Director, Dow Carter Ltd
Favourite LARTs #7: The Big Scary Angle Grinder(tm)
"I was playing poker the other night ... with Tarot cards. I got a full
house and four people died."
-- Steven Wright
Doug Freyubrger
>
> > But if they gave you a telephone and told you your extension number
> > was "3033" would you be miffed? Would you demand that they change
> > your phone number to "John"?
>
> Eh? Telephone numbers are traditionally, well numbers. I wouldn't
> expect a telephone number to be "John". I wouldn't expect it to be
> a nice shade of lilac. I'd expect it to be a number.
>
> Email addresses tend to be text. Hence "email address" and not
> "email number".
So what? Telephone numbers are listed in a database called a phone book
and people maintain there own roladex lists. E-mail addresses are listed
in a database called LDAP/ActiveDirectory/NIS/etc and people maintain
their own contact lists. Since e-mail addresses by default have the full
name associated with them, it makes storing them in your contact list
much easier than maintaining a roladex by hand.