On 5/19/08 7:46 PM, Michael Torrie [EMAIL PROTECTED] wrote:
Daniel Fussell wrote:
Michael Torrie wrote:
We're just using RHEL5's built-in xen support, which is decent, but not
as great as vmware server. or commercial xen for management and failover.
I was trying to configure Linux-HA a
I work with a research team that has a problem with version control. By the
time they're 3 weeks into a project, there is usually somewhere between 3-8
versions spawned from the original data set but now with different changes
made, and rarely does anyone remember what version had what changes
What version control system are they using currently? Windows has a really
great GUI for subversion, TortoiseSVN, http://tortoisesvn.tigris.org...all
the windows suckers here at work use it.
-Brandon
On Tue, May 20, 2008 at 9:08 AM, Jonathan Allred [EMAIL PROTECTED]
wrote:
I work with a
On Monday 19 May 2008 06:30:30 pm Stuart Jansen wrote:
Like everything, they're a tool that can be abused, but shouldn't be
summarily discounted as a poor man's way of programming.
Let me put it this way: I haven't found a use case where triggers
offered better performance *and* provided
On Monday 19 May 2008 11:33:32 pm Gabriel Gunderson wrote:
On Mon, 2008-05-19 at 14:14 -0600, Alberto Treviño wrote:
Sorry to be so picky. PostgreSQL is a very good database. So is
MySQL. Unfortunately, MySQL keeps getting a bad reputation from
people who don't keep up on its development.
AJ ONeal wrote:
Onboard controller. It doesn't recognize my ethernet card either (but
surprisingly did detect my wireless).
I recently tried to cure some winrot to play TF2 on my old socket 740
board (3.5 years, time to upgrade). Not only did it take a lot of BIOS
tinkering to turn on my
On Monday 19 May 2008 07:09:18 pm Daniel Fussell wrote:
I do have a SAN and fiber channel switch and have been wanting to use
it for xen (or DRBD if it's stable enough), but someone in my
department has heard rumors that vm's are IO bound. So it hasn't
gone far from there. I know vmware can
On Tue, May 20, 2008 at 9:08 AM, Jonathan Allred [EMAIL PROTECTED] wrote:
I work with a research team that has a problem with version control. By the
time they're 3 weeks into a project, there is usually somewhere between 3-8
versions spawned from the original data set but now with different
On Monday 19 May 2008 08:08:23 pm Stuart Jansen wrote:
Because the vanilla
kernel lags qumranet source, and distros ship old kernels, I have no
idea how many of these improvements have made it into the distros.
I compile for rapidly changing features, but those improvements
include better
On Tue, 20 May 2008 10:06:26 -0600, Alberto Treviño [EMAIL PROTECTED]
said:
Let me put it this way: I haven't found a use case where triggers
offered better performance *and* provided better data integrity over:
- Good database design
- Proper column definitions (type, default values, etc.)
On Tue, May 20, 2008 at 10:22 AM, Scott K. [EMAIL PROTECTED] wrote:
I recently tried to cure some winrot to play TF2 on my old socket 740
TF2 for the win!
~ Nathan
BYU Unix Users Group
http://uug.byu.edu/
The opinions expressed in this message are the responsibility of
Stuart Jansen wrote:
Like everything, they're a tool that can be abused, but shouldn't be
summarily discounted as a poor man's way of programming.
In fact, the entire debate over where the business logic should be is
worth reopening.
Web applications of late have kind of helped push the
Alex Esplin wrote:
On Tue, May 20, 2008 at 9:08 AM, Jonathan Allred [EMAIL PROTECTED] wrote:
I work with a research team that has a problem with version control. By the
time they're 3 weeks into a project, there is usually somewhere between 3-8
versions spawned from the original data set but
On Tuesday 20 May 2008 10:40:32 am Jonathan Ellis wrote:
Then you realize that this is also good practice even in a single
language.
Can you provide an example?
--
Alberto Treviño
[EMAIL PROTECTED]
Testing Center
Brigham Young University
BYU Unix Users Group
I know I have created a web application using php and connecting to a mysql
backend and just did it as a project for someone. I don't maintain the code
so if they start programming stuff for it the triggers and such will help
maintain integritybut I guess that could also be helped by a well
Alberto Treviño wrote:
On Tuesday 20 May 2008 10:40:32 am Jonathan Ellis wrote:
Then you realize that this is also good practice even in a single
language.
Can you provide an example?
In a roundabout way. A web app that needs a non-web frontend. Or how
about a conventional app that needs
One thing you are missing is that a lot of web apps now are RESTful in that
they provide an standard API to access their services. This way the model
and business logic stay in the framework but you are not locked into an
specific language at all. Personally I prefer the model and business logic
On Tue, 2008-05-20 at 11:26 -0600, Alberto Treviño wrote:
On Tuesday 20 May 2008 10:40:32 am Jonathan Ellis wrote:
Then you realize that this is also good practice even in a single
language.
Can you provide an example?
My predecessors wrote the system in COBOL. Then we decided to switch
On Tue, 2008-05-20 at 11:59 -0600, Adrian Madrid wrote:
One thing you are missing is that a lot of web apps now are RESTful in
that they provide an standard API to access their services. This way
the model and business logic stay in the framework but you are not
locked into an specific
Adrian Madrid wrote:
One thing you are missing is that a lot of web apps now are RESTful in that
they provide an standard API to access their services. This way the model
and business logic stay in the framework but you are not locked into an
specific language at all. Personally I prefer the
On Tuesday 20 May 2008 12:03:14 pm Stuart Jansen wrote:
On Tue, 2008-05-20 at 11:26 -0600, Alberto Treviño wrote:
On Tuesday 20 May 2008 10:40:32 am Jonathan Ellis wrote:
Then you realize that this is also good practice even in a single
language.
Can you provide an example?
snip
I
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Jonathan Allred
Sent: Tuesday, May 20, 2008 9:08 AM
To: uug-list@uug.byu.edu
Subject: [uug] Version Control System for Non-techies
I work with a research team that has a problem with version
On Tue, 2008-05-20 at 12:14 -0600, Michael Torrie wrote:
Stuarts hypothetical example is a great argument against your
preference, especially if you were working on a mission critical
application.
But let's not start acting like J2EE programmers. Not every system is
going to last 5+ years. Not
Alberto Treviño wrote:
On Tuesday 20 May 2008 12:03:14 pm Stuart Jansen wrote:
On Tue, 2008-05-20 at 11:26 -0600, Alberto Treviño wrote:
On Tuesday 20 May 2008 10:40:32 am Jonathan Ellis wrote:
Then you realize that this is also good practice even in a single
language.
Can you provide an
On Tue, 2008-05-20 at 12:30 -0600, Brian Phillips wrote:
Version control software is a bit too much overhead for the research
environment in my opinion.
Subversion, yes it might be. But have you tried distributed revision
control? I personally prefer git because of git rebase -i and a few
other
On Tue, 20 May 2008 12:30:37 -0600, Alberto Treviño [EMAIL PROTECTED]
said:
I still didn't see a specific example. What I heard were great cases
for stored procedures and proper foreign key constraints. I still
don't see a good case for triggers specifically.
Triggers are to databases what
On Tue, May 20, 2008 at 12:11 PM, Stuart Jansen [EMAIL PROTECTED]
wrote:
RESTful design is also of limited value when working with legacy code
that wasn't designed to take advantage of it. That doesn't mean it
shouldn't be used, it just means that triggers shouldn't be summarily
rejected.
On Tue, 20 May 2008 12:51:47 -0600, Adrian Madrid [EMAIL PROTECTED]
said:
Having the language is half the battle. The environment that language
runs
is as important. Do I get access to all the libraries, testing
environments,
distributed/scalability options?
Short answer: yes, if you can
On Tue, 2008-05-20 at 12:55 -0600, Adrian Madrid wrote:
I'm of the opposite opinion. I have used web services as the glue that
ties legacy systems to other more current systems. I gain all the
language/environment neutrality and I get to do it in my
language/environment of choice. I might not
On Tue, 2008-05-20 at 12:30 -0600, Alberto Treviño wrote:
I still didn't see a specific example. What I heard were great cases
for stored procedures and proper foreign key constraints. I still
don't see a good case for triggers specifically.
Okay, here's a couple of specific examples:
1)
If the only gateway is SQL then sure there is no other way. Now, if we are
talking web apps even if you don't have control over anything but login in
then you can scrape and build your gateway restfully. Slow and painful but
doable.
AEM
On Tue, May 20, 2008 at 1:06 PM, Stuart Jansen [EMAIL
On Tuesday 20 May 2008 12:54:58 pm Jonathan Ellis wrote:
(I'm not sure why I
didn't mark the keys in clan_members ON DELETE CASCADE. Probably
because I didn't know about that when I wrote it, a lot of CB code
dates back to 2001 when I didn't really know what I was doing.)
Exactly. ON DELETE
On Tue, 2008-05-20 at 13:11 -0600, Adrian Madrid wrote:
If the only gateway is SQL then sure there is no other way. Now, if we
are talking web apps even if you don't have control over anything but
login in then you can scrape and build your gateway restfully. Slow
and painful but doable.
So
On Tuesday 20 May 2008 12:36:11 pm Stuart Jansen wrote:
But let's not start acting like J2EE programmers. Not every system is
going to last 5+ years. Not every system needs to be carefully
designed to adapt to changing requirements. Sometimes the smart
solution is a design that is cheap and
On Tue, 2008-05-20 at 13:11 -0600, Adrian Madrid wrote:
If the only gateway is SQL then sure there is no other way. Now, if we
are talking web apps even if you don't have control over anything but
login in then you can scrape and build your gateway restfully. Slow
and painful but doable.
Oh,
On Tue, May 20, 2008 at 12:59 PM, Jonathan Ellis
[EMAIL PROTECTED] wrote:
On Tue, 20 May 2008 12:51:47 -0600, Adrian Madrid [EMAIL PROTECTED]
said:
Having the language is half the battle. The environment that language
runs
is as important. Do I get access to all the libraries, testing
On Tuesday 20 May 2008 01:11:38 pm Stuart Jansen wrote:
Okay, here's a couple of specific examples:
1) Auditing. On update insert records into a history table. It
shouldn't be possible to bypass auditing by ignore the API.
Thank you! You win the prize!
2) Updateable views, while some DBs
On Tuesday 20 May 2008 01:19:37 pm Adrian Madrid wrote:
I believe this whole thread is going a lot of places
Isn't that the truth! It has brought forth the religious battles of
MyDB over YourDB and myProgrammingStyle vs your_programming_style.
Should be continue now to KDE vs. Gnome, or VI
Blanket rules make no sense at all. So let me explain my case. Let's say I
need to work with some legacy web app named Gizmo. I have access to the code
so I can figure out all the DB calls. The requirements are that several
systems from several languages/environments could access the information
On Tue, 2008-05-20 at 13:19 -0600, Adrian Madrid wrote:
I believe this whole thread is going a lot of places but I got in
where the business logic should reside part. In that sense I believe
my shared-nothing attitude is not orthogonal but an important
principle for me to make these kind of
On Tue, 2008-05-20 at 13:28 -0600, Adrian Madrid wrote:
Blanket rules make no sense at all.
*snip*
Does that make sense?
Sure, I just wanted to find out where you stood on the whole triggers
are evil subject. Scraping can be just as brittle as triggers, and
takes a whole lot longer to
Sure, I was just getting into the busines logic residence dilemma.
On Tue, May 20, 2008 at 1:33 PM, Stuart Jansen [EMAIL PROTECTED] wrote:
On Tue, 2008-05-20 at 13:28 -0600, Adrian Madrid wrote:
Blanket rules make no sense at all.
*snip*
Does that make sense?
Sure, I just wanted to find
Stuart Jansen wrote:
Subversion, yes it might be. But have you tried distributed revision
control? I personally prefer git because of git rebase -i and a few
other features, but bzr might be a better starting point.
I was actually going to suggest Git, too, except that the OP seemed
rather
On Tue, 2008-05-20 at 13:26 -0600, Alberto Treviño wrote:
Should be continue now to KDE vs. Gnome, or VI vs. Emacs or Linux vs.
UNIX? (Just kidding!)
Thank you, I much prefer playing the role of rabid fan boy over
reasonable professional.
Vim, 'cause vi and emacs are both wrong. (But emacs
On Tue, 20 May 2008 13:12:28 -0600, Alberto Treviño [EMAIL PROTECTED]
said:
On Tuesday 20 May 2008 12:54:58 pm Jonathan Ellis wrote:
(I'm not sure why I
didn't mark the keys in clan_members ON DELETE CASCADE. Probably
because I didn't know about that when I wrote it, a lot of CB code
On Tue, 20 May 2008 13:19:37 -0600, Adrian Madrid [EMAIL PROTECTED]
said:
Can I control how many resources it will use? From which servers? Can I
replicate the environment and test it thoroughly? My point is that with
_normal_ ruby/python/etc deployments I can control and test every little
As I stated before I haven't done it with PG so I cannot attest if it is
possible and I was asking for information. You asked what I meant and I
answered. Did you read my post in context? Can you answer the question?
Can I control how many resources it will use? From which servers? Can I
Alberto Treviño wrote:
Let me put it this way: I haven't found a use case where triggers
offered better performance *and* provided better data integrity...
To answer the performance question using a real-world optimization I did
a couple months ago:
Suppose you have a table product_licenses,
Stuart Jansen wrote:
Maddog 'cause rms and esr are both scary
Over 'cause under is just stupid
Less 'cause less is more than more.
--Dave
BYU Unix Users Group
http://uug.byu.edu/
The opinions expressed in this message are the responsibility of their
author. They are
Alberto Treviño wrote:
On Tuesday 20 May 2008 02:02:12 pm Jonathan Ellis wrote:
No, you misread my post.
clan_*members* could have used a FK but clan_*participants* needed a
trigger.
You are right. I didn't understand your post because you didn't give me
the whole schema. I don't know
On Sun, May 18, 2008 at 07:28:04PM -0500, Mark Gardner wrote:
On Fri, May 16, 2008 at 7:45 PM, Adam Johnson [EMAIL PROTECTED] wrote:
Hi Unix User Group. I'm going to be a new freshman in Fall. I'm an
avid Linux user(Ubuntu 7.10 right now). I've been looking around the
IT department's page
Thanks!
Christijan
Web Developer
BYU Graduate Studies
(801) 422-5658
[EMAIL PROTECTED]
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Chris Carey
Sent: Tuesday, May 20, 2008 4:02 PM
To: [EMAIL PROTECTED]
Subject: Re: [uug] Ubuntu and Synergy
known bug
Anyway, it sounded like a natural fit since the OP mentioned that the
collaborators tend to go half a dozen directions from the starting
point. Git would allow them to pick the fruitful branches and merge if
they want, while discarding the ones they don't care about. And in the
Mark Gardner wrote:
But does it have a windows GUI?
Well, it does have a web client, which I've never used so I can't speak
to its quality. But Git is at its best when everyone has a local copy
of the repo, which I guess would mean everyone installing Apache (and
probably Cygwin) on their
Alberto Treviño wrote:
On Tuesday 20 May 2008 01:19:37 pm Adrian Madrid wrote:
I believe this whole thread is going a lot of places
Isn't that the truth! It has brought forth the religious battles of
MyDB over YourDB and myProgrammingStyle vs your_programming_style.
Should be
If you'd like to run for a position in the club, please send an email
to [EMAIL PROTECTED] tonight or early tomorrow morning. There's
been a few people respond already but we'll probably need more help.
Thanks,
Peter
BYU Unix Users Group
http://uug.byu.edu/
The
On Tue, May 20, 2008 at 11:50 AM, Stuart Jansen [EMAIL PROTECTED] wrote:
On Tue, 2008-05-20 at 12:30 -0600, Brian Phillips wrote:
Version control software is a bit too much overhead for the research
environment in my opinion.
Subversion, yes it might be. But have you tried distributed
57 matches
Mail list logo