I have also read Collective Intelligence. I think I enjoyed it significantly more than the Slashdot reviewer. Here is my review:
~~~~
Have you ever wondered how:
* Google comes up with its search results
* Amazon recommends you books/movies/music
* spam filters decide good from bad
Well, Toby Segaran not only explains these topics and more in Collective Intelligence, but he does so in a way accessible to software developers that haven't worked on machine-learning problems before. He even provides working Python code for all the algorithms.
Oh, and Collective Intelligence reads incredibly well. I could not wait to get home and get back to it -- and when I went in to work the next morning, I usually had a new idea or two of how to improve our software. I also started implementing the most important examples in Groovy to make sure I got it.
If you are a Senior Software Engineer or "better," this is a must-read. Proper application of the algorithms in this book are a great way to simplify your system and avoid getting nickel-and-dimed to death with new ways to prioritize/categorize/slice-and-dice your domain data.
I've done [Java] code reviews with and without static analysis tools.
If you don't have static analysis tools in place, the review tends to spend its time on finding and discussing little bugs. You'll still find the big ones, but the room is likely to be so worked up (or tuned out) by the discussion of "basics" to some and "nit-picking" to others that the discussion of important stuff is difficult to keep constructive.
If you do have static analysis tools in place, you can set the expectation that the reports of the static analysis tools are clean prior to the review. This allows the discussion to go straight to the important stuff. Maven's reporting features (plugins) are the number one reason I use it. My standard reports:
Changelog - retrieves list of recent changes from your cm system (CVS, Subversion, etc)/li>
JXR (generates an HTML version of the source that other plugins can reference -- this is great if a report highlights a problem, you can browse straight to it)/li>
Static analysis tools are one of the great resources of the Java language.
SOX is all about making sure you have processes and policy to keep your business financially safe (disaster recovery, access controls, financial controls, etc.) and auditing to make sure you're following your own process/policy.
And I agree. However I don't agree that:
The truth is that those controls can be anything you want them to be. Don't want to do backups? Just document why and be ready to explain it to the auditor. You are still in "compliance".
I think it is unlikely that you could ever explain to a non-conspiring auditor that your backup policy and process is adequate. SOX 404 definitely leaves a lot open to interpretation, but it's not *totally* arbitrary. There's just so many degrees of freedom that it often seems that way.
Isn't it terribly ironic that the audit industry has benefited *massively* from all the accounting scandals?
You're quite right. The average for this dataset is essentially meaningless because the data is not normally distributed (Anderson Darling p-value is <0.005). Minitab says the critical issue resolution process might follow an exponential or gamma distribution.
The median would be a better measure of the location of the "center" of this dataset.
The median is: 18 days Lower bound of 95% confidence interval for median: 11.137 Lower bound of 95% confidence interval for median: 21.621
This means that you can be 95% confident that the median time to critical issue resolution is between 11 and 22 days.
Total cost in mass production, not counting marketing, warranty, and other post-production costs: Probably under $50, not counting monitor. Monitor: $50 for new 15" CRT, significantly more for LCD screen.
Been there, researched that. Commercial thin client solutions cost more than $100; they typically start around $200 for the bare-bones models. I'm not sure what the cause is (I'm guessing volume), but you can purchase a low-end desktop PC from Dell for less than you can purchase a thin client suitable for the same tasks, even in volume.
Neoware is a leader in thin-client computing (along with being really swell guys) and is who IBM partners with when selling "thin-client" solutions. HP also has some thin-client products. They are as expensive as Neoware, though they do have really cool integration with the Altiris management software. However, the Altiris software works with Windows and Linux on standard PCs as well, so you can get the managability of thin clients using "desktop" hardware + Altiris. That was my recommendation to management, anyway.
Have you ever read The Promise of Sleep? I don't have anything that would classify as a sleep disorder (just a part of the "too busy" crowd), but I still found this book useful. It's written by Dr. Dement, who is one of the pioneers of sleep research and disorder treatment. The book has helpful advice on how to get in touch with doctors who understand and can treat sleep disorders.
You can also find it at Barnes & Noble. Should be less than $15.
I would have to say Mr. Cuban is more of a businessman than anything...what kind of tech enthusiast brags about his new HP Media Center PC?
You're probably right, Mark is more of a businessman than a technologist. He might be examining his target market (the huge numbers of people who buy consumer-grade computers) to see what kind of stuff is viable or could be viable within 5 years. He's clearly doing a lot of thinking about "the next 5 years", talks about distributing HD content on hard drives, and he specifically mentions it having a 160GB removable hard drive.
Holy cow. When I started reading your comment I started thinking about my days on the probe-perf list and the sweet turbo-calc spreadsheets that "hyc," Ross, Dan, and the guy in Kentucky published. And here you are. Small world. That was a good list.
I recall that you did set up the turbo, didn't you? Do you still have the car? I succeeded in hacking a TC into my 626, but I removed the TC and sold the 626 for a Eclipse GSX soon after completing the project. Then I bought a house, and that was the end of fun with cars.
The way we approach the "correct" JVM issue in our deployment environment is through Java Web Start. Via JWS, we can actually deploy the correct (the one we tested/qualified our app on) JVM to the client machine along with any application updates, when the application starts. If no updates are required or the JWS web server is down, the application launches from a cached local copy. JWS works extremely well for internal applications (and Open Source ones like ArgoUML). We have been using JWS for this production critical application for over a year and a half with no problems at all.
JWS has been brilliant for our uses. You might give it a look. Here is Sun's page on it: http://java.sun.com/products/javawebstart/
AC, This was clearly a design decision and if you prefer this style of monitoring, then I'd suggest Big Brother. For my environment, Nagios made the correct choice. If you are monitoring many applications (many > 100), then with a model that pushes events to the monitoring system, you will (probably) end up with a distributed configuration nightmare.
That said, I think you could probably hack a Nagios setup to do what you want with its distributed monitoring features. I.e., you could write your custom monitoring app to implement the interface that Nagios uses for satellite monitoring instances and then configure Nagios to use your custom monitoring app as a satellite. But I have not tried/done this, so I could be wrong, wrong, wrong.
However, if you were to state that you are so confident that a constant temperature of 'pleasant' wouldn't damage the servers that you will offer to pay for any overheating damage done out of your own salary (minus the cost savings realized by using less air conditioning), they might be willing to turn down the AC.
I'm all for accountability, but proposing they take any losses out of his salary is a bit extreme. It's not like they're going to give him all the money they save if everything works out ok. A simple return-on-investement analysis should do the trick for a management team that is responsive to reasonable employee input. If that doesn't work, he's got bigger problems anyway.
1) Never drive a car that is worth more than two months gross pay.
Almost a 'check' here. My car is worth about 3 months of my current pay. Of course, I'm a car guy, so that influenced my choice a great deal. The interesting thing I did recently was to 'refinance' the car loan (9.25% ream-the-new-grad-rate) via a balance transfer to my credit card (banks won't touch a $5k used-car loan). Gasp! Balance on Credit Card! Well, this transfer was done at 2.9% *for the life of the balance*. If you use this strategy, make sure you are not currently carrying a balance (probably a 7-14% rate) on the card since your payments will go to the lowest-rate part of your balance first. After all the cash flows were computed, turns out I'm saving $800 here.
2) DON'T BUY A HOUSE!
Did it! I know, you're beginning to think I'm stupid. Of course, after factoring in the mortgage interest deductions effect on my taxes, my 3br house is less on a yearly basis than my 2br apartment was. And the property value is increasing due to my location, location, location.
My greatest concern here is that I bought without having the cash in the bank to get me through a layoff without cannibalizing the 401(k). Fixing that now.
You could buy 2 acres about 30 miles out of town for $10,000 and put a mobile home on it (you can build a house on it later).
This is IMHO, but there's a difference between 'living' and 'not dead'. 30mi from work? Spending 1.5hrs in the car every day (assuming 45 min each way) and accumulating 15k mi/yr on the car doesn't sound like a wise investment of resources. If you've got that much extra time, get a part-time job consulting or something.
3) Get out of debt.
My parents gave me the gift of a debt-free education. This is huge. Of course, they'll probably get it all back in about 15-20yrs, but I will be more capable of paying at that time. I will make sure to do the same for my (non-existent) kids. Aside from the house and the car, I've managed to steer clear of debts. Carrying a CC balance makes my skin crawl.
Aside from all that, I've found that the rate at which I spend money is directly influenced by a couple of non-obvious (to me) factors:
1. Ease of spending: A couple of weeks ago, I (thought) lost my wallet. Turned out it was in a pair of shorts in my dresser, but whatever. However, what happened is that I had no credit or debit cards, and no easy way of getting more cash (write check to friend, have friend go to ATM). The credit card companies were amazingly slow at replacing them (which surprised the hell out of me). The result was that I spent ~$300/mo instead of the the usual ~$600/mo. What did that teach me? I really only need $300 worth of stuff to cover me in a month because that's what I spend when it's not a huge PITA. I'm going to shoot for $400/mo and the other $200 will go into the rainy-day fund.
2. Boredom: Turns out it's a lot more expensive to be entertained than it is to entertain yourself. So, I decided to invest in myself (and I don't mean cosmetic surgery). I read a lot more and I've discovered the libraries around here are great. I'm training for a marathon. I'm in a $FAV_SPORT league. All activities that take (large amounts of) time, don't require a lot of equipment, and are generally self-improving, moreso than Die Hard II, anyway. God bless John McClane.
I found this regarding SnapFS, which is a layer that can sit between the Linux VFS and a journaled filesystem to perform (not just) versioning. Copied from LWN.net:
Whenever a file is to be modified, and its contents must be preserved in a snapshot, SnapFS creates a new inode in the filesystem to hold the snapshot version. An extended attribute which points to the snapshot inode is then attached to the visible version of the file. The actual blocks of the file are shared between the current file and the snapshot until they are changed;
at that point the SnapFS "copy on write" mechanism makes copies of the affected blocks . Snapshots are thus relatively efficient in their use of storage, especially in situations where only parts of files are changed. For example, a snapshot of that huge web server log file, which is only appended to, does not duplicate the log entries that are shared between the current and archived versions.
Unfortunately, I think this project may be stalled as noted in the developer's project notes.
I still don't konw what OpenVMS does as I couldn't determine that after over an hour with Google. However, looking at the semantics of the DELETE and PURGE commands it looks like you can get remove arbitrary versions of files. So, I guess each version is a complete copy (or can become one when you delete earlier versions).
Ok. Ziti and Sopranos time. Let's hear it for season 3!
"It never overwrites old files..." Many like this feature: by putting those hooks in at the filesystem level, all commands automatically inherit file versioning. When you're certain you don't need the old versions any longer, you can clean up with a single command. And, finally, if you really don't like it, you can turn it off.
Do you know how the versioning is implemented? I am a 'casual' user of VMS 7.2-1/7.3 at work, and this I think the automatic file versioning is a really cool feature. However, I've always wondered if I make a small change to a large file, does it copy the whole file? I'm guessing there are some configurable limits on file size for what gets versioned and what doesn't. I suppose you could use some sort of diff/patch implementation, but that would get a bit messy if you have (to borrow an example from elsewhere in this thread) foo;1 foo;2 foo;3 and delete foo;2.
Another feature I really like about VMS is the performance reporting tools that are built (??) into it. Every time I run load tests without telling my admin, he starts sending me graphs about what I did to the machines and do I know anything about it, plus if I can ask him for data from 2 months ago, and he's likely to have it.
Stephen
P.S. I say 'casual' because I just can't get into the 'set def mylogical:[sub_dir]' sort of stuff that our applications require.
What you're saying is true, but only moreso in a corporate environment. In the large corp that I work at, most people use: 1. Email 2. An office productivity suite 3. Web browser 4. One or two job-specific off-the-wall/custom applications
Now, the PC hardware they do this on is worth about ~$400 if it's a desktop and ~$800 if it's a laptop. Corporate IT departments are beginning to wonder "Why am I increasing the cost of my solutions by 50-100% per-user (and increasingly, per-year) by using MS stuff?". That doesn't even account for the money spent keeping the systems humming, which evidently is >$1000 per user, per year for the MS platforms at my environment. An alternative that is emerging is Linux booting off of either the local disk or a remote image which can be centrally managed. Run Mozilla/Evolution, Mozilla, and OpenOffice for 1, 2, and 3. Use wine or a Windows Terminal Server (quite a good product, actually) via rdesktop (this is an MS EULA violation, I think) for #4.
Regards, Stephen
Re:Ctrl-Tab Analogue in Mozilla's Tabbed Browsing?
on
A First Look at Netscape 7
·
· Score: 5, Informative
My manager (a genuine SW fan) organized a trip to the theater to see SW:AotC on opening day. The company didn't pay for the tickets, but everyone was excited to: a) see Star Wars before everyone else did (including their kids) b) spend some time together where there was no pressure to perform
As an employee of a company that has seen 4-5 layoffs over the past year and a half, and a member of a group that recently completed a very stressful project, I can say this trip was the best thing anyone has done for my group's morale and general stress level in at least six months (and it didn't really cost the company anything). Of course, that probably means we should have been doing stuff like this all along.
Stephen
P.S. For those about to say that it cost the company 20 man hours for my group to go see this movie, just take it out of my 'overtime account'. There's about 500 unpaid hours in there already for this year.
man_ls said: But it wouldn't help with something like a web project where multiple sites all had the same data on them, changing that one file would change them all.
It depends on whether you want them to share some files or not. You may want to have all sites share certain templates, for instance.
I'll give you a practical example of its usage. Say you're a Java developer and you need to run multiple instances of your application, but under different configurations. You can set up a filesystem like the following:
directory containing the core libraries for the current version of your application: /apps/myapp-1.1
a symbolic link to the 'current' version of your application's core libraries:
/apps/current-version ->/app/myapp-1.1
directories which provide the special sauce/configuration for whatever the app needs to run with: /apps/instance-1 .../apps/instance-n
Then you can make set your path/classpath to the following:
$CLASSPATH=/apps/instance-1:/apps/current-version
This classpath will allow you to upgrade the core libraries of all instances of your application at once with a command sequence like:
rm current-version
ln -s myapp1.2 current-version
restart myapp
Or, perhaps more importantly, it will let you verify that all instances are running the correct version very easily (ls -l current-version).
That's something that you can't do on Win2k (don't know about XP, but since there's no server version, I'm not interested), even though it's incredibly useful to people running server applications. One of my MS admin friends at work did tell me there was a way to perform linking of directories with a utility in the latest Win2k Server Resource Kit but it certainly isn't installed by default and MS charges for the resource kit, so it's yet-another-thing-to-license and you can't expect 'normal' MS admins to be familiar with it.
On your webpage (which is great, btw), you describe seeing sediment in your coolant. Water Wetter will produce sediment as well. I believe it's either grey in color and flaky or copper in color and flaky. I can't remember which. It usually freaks car people out when they find it in their coolant and think they've blown a head gasket or something. It should not be too hard to verify the exact nature of the sediment.
I know this doesn't answer your question, but you might try Dia. It has a GIMP-style (frameless, is it?) drawing editor. I have found I prefer it to Visio for things like UML (which I think they have done a fabulous job with -- better than ArgoUML IMO). I haven't used it for things other than UML, so I won't talk about those.
The world's most "innovative", "enterprise-ready" OS, developed in the world's most prestigious and ornate cathederal -- won't run if they remove the web browser component -- even if it costs the company millions of dollars per day?
Steve and [Jim] Allchin are blowing smoke up everyone's ass whenever they talk to people outside of MS.
If it's the first option, I'd fire the guy running the design reviews. If it's the second option, I'd throw Ballmer in jail because he lied in a deposition.
On a side note, Ballmer goes on for a bit talking about how it's impossible to document the internal apis so as to make everything modular. Why? I thought that was one of the ideas behind [D]COM[+]. It might even spur a bit of real innovation in how to document and modularize system components like authentication, (distributed) file-trading, window management, etc.
I'm glad to hear that 4 of the top 5 mobile phone vendors aren't taking this bait, but I cringed when I read the following:
These include the capacity to offer multimedia text and picture messaging, or simpler access to corporate email or common business software programs, to mention just a few of the growing ranges of functions from such higher-powered phones.
Where:
'corp email' == Outlook && 'common biz software' == Office
Which got me thinking, that yes, that would be nice, I wonder why the Java developers haven't done that yet as well with the phone dev kits that are already shipping. The answer, of course, is that none of those protocols are open and thus can't.
Did the proposed anti-trust settlement address this point? This is *classic* Microsoft market-leveraging behaviour. Without this point of leverage, I would expect this initiative to fail because battery life is too precious to waste on the inevitably large memory/processor requirements of this OS. (Of course, I'm assuming bloated code here, does anyone have any idea what the resource requirements are?)
I'd expect that it'd be easier/cheaper/more natural from a design standpoint to allocate a bit more fuel to de-orbit the satellite properly than to tack on a self-destruction system (however simple) because of additional explosive/control system/weight requirements.
Of course, what would happen then is the management/owners would run the satellite until it was absolutely, positively out of fuel and then we'd be back in the same boat; debris be damned. Better ask people to build satellites with modular designs and throw some explosive bolts in there so we don't tempt fate.
Slashdot Mirror
I have also read Collective Intelligence. I think I enjoyed it significantly more than the Slashdot reviewer. Here is my review:
~~~~
Have you ever wondered how:
* Google comes up with its search results
* Amazon recommends you books/movies/music
* spam filters decide good from bad
Well, Toby Segaran not only explains these topics and more in Collective Intelligence, but he does so in a way accessible to software developers that haven't worked on machine-learning problems before. He even provides working Python code for all the algorithms.
Oh, and Collective Intelligence reads incredibly well. I could not wait to get home and get back to it -- and when I went in to work the next morning, I usually had a new idea or two of how to improve our software. I also started implementing the most important examples in Groovy to make sure I got it.
If you are a Senior Software Engineer or "better," this is a must-read. Proper application of the algorithms in this book are a great way to simplify your system and avoid getting nickel-and-dimed to death with new ways to prioritize/categorize/slice-and-dice your domain data.
You are absolutely correct.
I've done [Java] code reviews with and without static analysis tools.
If you don't have static analysis tools in place, the review tends to spend its time on finding and discussing little bugs. You'll still find the big ones, but the room is likely to be so worked up (or tuned out) by the discussion of "basics" to some and "nit-picking" to others that the discussion of important stuff is difficult to keep constructive.
If you do have static analysis tools in place, you can set the expectation that the reports of the static analysis tools are clean prior to the review. This allows the discussion to go straight to the important stuff. Maven's reporting features (plugins) are the number one reason I use it. My standard reports:
Static analysis tools are one of the great resources of the Java language.
Stephen
And I agree. However I don't agree that:
I think it is unlikely that you could ever explain to a non-conspiring auditor that your backup policy and process is adequate. SOX 404 definitely leaves a lot open to interpretation, but it's not *totally* arbitrary. There's just so many degrees of freedom that it often seems that way.
Isn't it terribly ironic that the audit industry has benefited *massively* from all the accounting scandals?
StephenYou're quite right. The average for this dataset is essentially meaningless because the data is not normally distributed (Anderson Darling p-value is <0.005). Minitab says the critical issue resolution process might follow an exponential or gamma distribution.
The median would be a better measure of the location of the "center" of this dataset.
The median is: 18 days
Lower bound of 95% confidence interval for median: 11.137
Lower bound of 95% confidence interval for median: 21.621
This means that you can be 95% confident that the median time to critical issue resolution is between 11 and 22 days.
Stephen
Been there, researched that. Commercial thin client solutions cost more than $100; they typically start around $200 for the bare-bones models. I'm not sure what the cause is (I'm guessing volume), but you can purchase a low-end desktop PC from Dell for less than you can purchase a thin client suitable for the same tasks, even in volume.
Check the prices for yourself:
http://www.neoware.com/thin-clients/capio_one.htm
Neoware is a leader in thin-client computing (along with being really swell guys) and is who IBM partners with when selling "thin-client" solutions. HP also has some thin-client products. They are as expensive as Neoware, though they do have really cool integration with the Altiris management software. However, the Altiris software works with Windows and Linux on standard PCs as well, so you can get the managability of thin clients using "desktop" hardware + Altiris. That was my recommendation to management, anyway.
Regards,
Stephen
You can also find it at Barnes & Noble. Should be less than $15.
You're probably right, Mark is more of a businessman than a technologist. He might be examining his target market (the huge numbers of people who buy consumer-grade computers) to see what kind of stuff is viable or could be viable within 5 years. He's clearly doing a lot of thinking about "the next 5 years", talks about distributing HD content on hard drives, and he specifically mentions it having a 160GB removable hard drive.
Stephen
Holy cow. When I started reading your comment I started thinking about my days on the probe-perf list and the sweet turbo-calc spreadsheets that "hyc," Ross, Dan, and the guy in Kentucky published. And here you are. Small world. That was a good list.
I recall that you did set up the turbo, didn't you? Do you still have the car? I succeeded in hacking a TC into my 626, but I removed the TC and sold the 626 for a Eclipse GSX soon after completing the project. Then I bought a house, and that was the end of fun with cars.
Stephen
The way we approach the "correct" JVM issue in our deployment environment is through Java Web Start. Via JWS, we can actually deploy the correct (the one we tested/qualified our app on) JVM to the client machine along with any application updates, when the application starts. If no updates are required or the JWS web server is down, the application launches from a cached local copy. JWS works extremely well for internal applications (and Open Source ones like ArgoUML). We have been using JWS for this production critical application for over a year and a half with no problems at all.
JWS has been brilliant for our uses. You might give it a look. Here is Sun's page on it:
http://java.sun.com/products/javawebstart/
Regards,
Stephen
AC,
This was clearly a design decision and if you prefer this style of monitoring, then I'd suggest Big Brother. For my environment, Nagios made the correct choice. If you are monitoring many applications (many > 100), then with a model that pushes events to the monitoring system, you will (probably) end up with a distributed configuration nightmare.
That said, I think you could probably hack a Nagios setup to do what you want with its distributed monitoring features. I.e., you could write your custom monitoring app to implement the interface that Nagios uses for satellite monitoring instances and then configure Nagios to use your custom monitoring app as a satellite. But I have not tried/done this, so I could be wrong, wrong, wrong.
Regards,
Stephen
Regards,
Stephen
Did it! I know, you're beginning to think I'm stupid. Of course, after factoring in the mortgage interest deductions effect on my taxes, my 3br house is less on a yearly basis than my 2br apartment was. And the property value is increasing due to my location, location, location. My greatest concern here is that I bought without having the cash in the bank to get me through a layoff without cannibalizing the 401(k). Fixing that now. This is IMHO, but there's a difference between 'living' and 'not dead'. 30mi from work? Spending 1.5hrs in the car every day (assuming 45 min each way) and accumulating 15k mi/yr on the car doesn't sound like a wise investment of resources. If you've got that much extra time, get a part-time job consulting or something. My parents gave me the gift of a debt-free education. This is huge. Of course, they'll probably get it all back in about 15-20yrs, but I will be more capable of paying at that time. I will make sure to do the same for my (non-existent) kids. Aside from the house and the car, I've managed to steer clear of debts. Carrying a CC balance makes my skin crawl.
Aside from all that, I've found that the rate at which I spend money is directly influenced by a couple of non-obvious (to me) factors:
1. Ease of spending: A couple of weeks ago, I (thought) lost my wallet. Turned out it was in a pair of shorts in my dresser, but whatever. However, what happened is that I had no credit or debit cards, and no easy way of getting more cash (write check to friend, have friend go to ATM). The credit card companies were amazingly slow at replacing them (which surprised the hell out of me). The result was that I spent ~$300/mo instead of the the usual ~$600/mo. What did that teach me? I really only need $300 worth of stuff to cover me in a month because that's what I spend when it's not a huge PITA. I'm going to shoot for $400/mo and the other $200 will go into the rainy-day fund.
2. Boredom: Turns out it's a lot more expensive to be entertained than it is to entertain yourself. So, I decided to invest in myself (and I don't mean cosmetic surgery). I read a lot more and I've discovered the libraries around here are great. I'm training for a marathon. I'm in a $FAV_SPORT league. All activities that take (large amounts of) time, don't require a lot of equipment, and are generally self-improving, moreso than Die Hard II, anyway. God bless John McClane.
Well, those are my thoughts.
Regards,
Stephen
I still don't konw what OpenVMS does as I couldn't determine that after over an hour with Google. However, looking at the semantics of the DELETE and PURGE commands it looks like you can get remove arbitrary versions of files. So, I guess each version is a complete copy (or can become one when you delete earlier versions).
Ok. Ziti and Sopranos time. Let's hear it for season 3!
Regards,
Stephen
Another feature I really like about VMS is the performance reporting tools that are built (??) into it. Every time I run load tests without telling my admin, he starts sending me graphs about what I did to the machines and do I know anything about it, plus if I can ask him for data from 2 months ago, and he's likely to have it.
Stephen
P.S. I say 'casual' because I just can't get into the 'set def mylogical:[sub_dir]' sort of stuff that our applications require.
Well, it looks like a guy named Colin Phipps is maintaining this source. You can reach him at:
cph [at] cph [dot] demon [dot] co [dot] uk.
If you have a more specific contact request, then let me know and I will try to find it.
Regards,
Stephen
What you're saying is true, but only moreso in a corporate environment. In the large corp that I work at, most people use:
1. Email
2. An office productivity suite
3. Web browser
4. One or two job-specific off-the-wall/custom applications
Now, the PC hardware they do this on is worth about ~$400 if it's a desktop and ~$800 if it's a laptop. Corporate IT departments are beginning to wonder "Why am I increasing the cost of my solutions by 50-100% per-user (and increasingly, per-year) by using MS stuff?". That doesn't even account for the money spent keeping the systems humming, which evidently is >$1000 per user, per year for the MS platforms at my environment. An alternative that is emerging is Linux booting off of either the local disk or a remote image which can be centrally managed. Run Mozilla/Evolution, Mozilla, and OpenOffice for 1, 2, and 3. Use wine or a Windows Terminal Server (quite a good product, actually) via rdesktop (this is an MS EULA violation, I think) for #4.
Regards,
Stephen
Try Ctrl+PgUp and Ctrl+PgDn. You can find other shortcuts here:m ozkeylist.html
http://www.mozilla.org/projects/ui/accessibility/
Of special interest (to me) are:
Ctrl+T - New tab with focus in location entry box
Ctrl+W - Close Tab
Regards,
Stephen
My manager (a genuine SW fan) organized a trip to the theater to see SW:AotC on opening day. The company didn't pay for the tickets, but everyone was excited to:
a) see Star Wars before everyone else did (including their kids)
b) spend some time together where there was no pressure to perform
As an employee of a company that has seen 4-5 layoffs over the past year and a half, and a member of a group that recently completed a very stressful project, I can say this trip was the best thing anyone has done for my group's morale and general stress level in at least six months (and it didn't really cost the company anything). Of course, that probably means we should have been doing stuff like this all along.
Stephen
P.S. For those about to say that it cost the company 20 man hours for my group to go see this movie, just take it out of my 'overtime account'. There's about 500 unpaid hours in there already for this year.
man_ls said:
/apps/myapp-1.1
/apps/current-version -> /app/myapp-1.1
/apps/instance-1
... /apps/instance-n
But it wouldn't help with something like a web project where multiple sites all had the same data on them, changing that one file would change them all.
It depends on whether you want them to share some files or not. You may want to have all sites share certain templates, for instance.
I'll give you a practical example of its usage. Say you're a Java developer and you need to run multiple instances of your application, but under different configurations. You can set up a filesystem like the following:
directory containing the core libraries for the current version of your application:
a symbolic link to the 'current' version of your application's core libraries:
directories which provide the special sauce/configuration for whatever the app needs to run with:
Then you can make set your path/classpath to the following:
$CLASSPATH=/apps/instance-1:/apps/current-version
This classpath will allow you to upgrade the core libraries of all instances of your application at once with a command sequence like:
rm current-version
ln -s myapp1.2 current-version
restart myapp
Or, perhaps more importantly, it will let you verify that all instances are running the correct version very easily (ls -l current-version).
That's something that you can't do on Win2k (don't know about XP, but since there's no server version, I'm not interested), even though it's incredibly useful to people running server applications. One of my MS admin friends at work did tell me there was a way to perform linking of directories with a utility in the latest Win2k Server Resource Kit but it certainly isn't installed by default and MS charges for the resource kit, so it's yet-another-thing-to-license and you can't expect 'normal' MS admins to be familiar with it.
Regards,
Stephen
On your webpage (which is great, btw), you describe seeing sediment in your coolant. Water Wetter will produce sediment as well. I believe it's either grey in color and flaky or copper in color and flaky. I can't remember which. It usually freaks car people out when they find it in their coolant and think they've blown a head gasket or something. It should not be too hard to verify the exact nature of the sediment.
Regards,
Stephen
I know this doesn't answer your question, but you might try Dia. It has a GIMP-style (frameless, is it?) drawing editor. I have found I prefer it to Visio for things like UML (which I think they have done a fabulous job with -- better than ArgoUML IMO). I haven't used it for things other than UML, so I won't talk about those.
Regards,
Stephen
If it's the first option, I'd fire the guy running the design reviews. If it's the second option, I'd throw Ballmer in jail because he lied in a deposition.
On a side note, Ballmer goes on for a bit talking about how it's impossible to document the internal apis so as to make everything modular. Why? I thought that was one of the ideas behind [D]COM[+]. It might even spur a bit of real innovation in how to document and modularize system components like authentication, (distributed) file-trading, window management, etc.
Regards,
Stephen
I'm glad to hear that 4 of the top 5 mobile phone vendors aren't taking this bait, but I cringed when I read the following:
These include the capacity to offer multimedia text and picture messaging, or simpler access to corporate email or common business software programs, to mention just a few of the growing ranges of functions from such higher-powered phones.
Where:
'corp email' == Outlook && 'common biz software' == Office
Which got me thinking, that yes, that would be nice, I wonder why the Java developers haven't done that yet as well with the phone dev kits that are already shipping. The answer, of course, is that none of those protocols are open and thus can't.
Did the proposed anti-trust settlement address this point? This is *classic* Microsoft market-leveraging behaviour. Without this point of leverage, I would expect this initiative to fail because battery life is too precious to waste on the inevitably large memory/processor requirements of this OS. (Of course, I'm assuming bloated code here, does anyone have any idea what the resource requirements are?)
Stephen
I'd expect that it'd be easier/cheaper/more natural from a design standpoint to allocate a bit more fuel to de-orbit the satellite properly than to tack on a self-destruction system (however simple) because of additional explosive/control system/weight requirements.
Of course, what would happen then is the management/owners would run the satellite until it was absolutely, positively out of fuel and then we'd be back in the same boat; debris be damned. Better ask people to build satellites with modular designs and throw some explosive bolts in there so we don't tempt fate.
Regards,
Stephen
I didn't expect to find a review of Duron v. P4, but Tom's did one: http://www.tomshardware.com/cpu/01q4/011116/index. html
Regards,
Stephen
Satisfied Duron 800 owner