Ya know, I was going to mod the parent up; but, I felt the need to express just how much I agree with him or her. I know that Linden is probably an excellent language for its niche; but, the number of languages in fairly common use seems to have exploded. I am all for the freedom of choice; and, I think that having a variety of tools that can be selected based on performance, flexibility or preference is wonderful.
My perception is that there is a optimal balance between choice and usefulness. If we were living in a world where the only languages were Assembler and BASIC programming would be hugely difficult and software would be suck because the tools were so mismatched to the application. On the other extreme, if there was a language created for every project, programming would be hugely difficult and software would suck because not only would the application have to be written, but the language as well.
Somewhere in the middle is the perfect balance between endless variety of languages and extreme limitations. My feeling is that if some of the effort by the very intelligent and skilled developers who are creating new languages were instead channeled into improving existing languages, the limitations of current tools would be greatly reduced and the experience of all those who know existing languages would be relevant to solving problems. If a developer has to devote time and effort learning a new language, basic syntax, code conventions, optimal design patterns, etc., then that developer is losing time actually writing useful code.
I have written "Hello World" or some generic "Music Store" type app learning:
BASIC
Pascal
HyperCard
Fortran
C
C++
Java
Perl
PHP
Python
DOS/CMD Batchfiles
Shell Scripting
Kixtart
Ruby
Visual Basic
C#
MSH/Monad/Microsoft's new shell
Lisp
Now, I am not a programmer by profession. I've taken courses and I like to write little programs for personal use; so, my knowledge of these languages is minimal. But, I have had to learn what I know so that I could use existing code or frameworks to accomplish what I wanted. The same requirements for redundant learning applies to those who write code on a professional level. If a highly skilled C++ programmer wants to contribute to a project written in MONO/C#, there is a learning curve. If he then wants to work on a project in Python, another learning curve.
I would be willing to bet that some projects miss out on a lot of contributed code because people who are willing and have the algorithmic skills to assist are simply unwilling to take the time to learn whatever language is used.
Sorry it this was less than readable. It is late and I am tired and hungry. Off to the kitchen...
I started my career in IT installing a NetWare server in a trucking company that had previously used Wyse terminals connected to an IBM big-iron box in the early 90s. During most of the 90s I worked on Novell NetWare networks in companies like Westinghouse, Bank of America, GTE, Marriott Intl., etc. It was only when Windows 2000 came out that the customers I worked with started leaving the NetWare fold. Even now, I just finished a gig with a municipality helping them finish their transition from NetWare to 2003 Server. And, in 2007, they still have not migrated away from GroupWise.
So, though NetWare is a legacy product, Novell is very well know, and, in many cases liked, by corporate IT managers. I have very fond memories of using tools like ZenWorks to roll out hundreds of Windows 98 clients with little hands-on effort while RIS and SMS were just a gleam in Mr. Balmer's eye. Though I have no data to support it, I would be willing to bet that Novell has a lot more brand recognition than Red Hat in the enterprise market.
I remember those kits that made Fieros look like Ferraris back in the 80s. I do not think that Ferrari's reputation suffered a bit; though, I do not think much of the schmucks that bought the kits.
As long as no one is selling the "skinned" phones as if they were legit iPhones, I'm not sure where the harm to Apple exists.
I was wondering if any legal history scholars happen to be slashdot regulars who could explain the parallels (or lack thereof) between Alcohol Prohibition and the current efforts to restrict digital content. It seems that there should be some relationship there, though I have nowhere near the education to analyze it myself.
Thanks.
I just went to RaLink's site; and, I'd buy a RT2800P based PCI NIC to test. The only problem is, I cannot find who retails these devices. It is hard to support these companies when it seems that noone will link to actual purchaseable products. Even google fails to come up with a vendor when searching for RT2800P.
When I go to some of the non-cooperative OEM sites, they provide links to retailers. And, I can find tons of stuff on newegg.com with the closed firmware hardware; but, nothing with open firmware.
Can someone point me to a US retailer that sells the products that this article suggests, please?
For what it's worth, this is my transition experience... I have been working as an IT professional first with NetWare on the server and DOS/Win9x clients and for the last few years in totally Windows environments. My current project is automating software distribution and maintenance for a municipal government with tools like SMS, RIS, etc.
In the 99-00 timeframe, I had downloaded and dual-booted RedHat's distro for a few months. After the novelty wore off, I ended up never using Linux, so, I eventually gave the experiment up. About 8-9 months ago, I decided to give it another go. My initial try was Gentoo, because it seemed sensible to use a source based distro for open-source software. Well, it was cool in a "taking apart the old mechanical clock" sort of way; but, it took forever to download and compile everything.
Now, I know that if I had been more careful about my approach it might have gone somewhat quicker; but, a multi-day OS install while googling on another PC was not what I wanted. So, I tried Fedora, Slackware, SuSE, Debian and Ubuntu. I really like both Debian and Ubuntu (a Debian derivative) because of the apt package management tools. I had difficulty with Debian because I do not have a wired ethernet connection to my desk; and, having to pick the PC, monitor and keyboard up and go to the router was not convenient, and, Debian does not have the non-free wireless drivers I need as part of their base installer CD.
Ubuntu, on the other hand, installed quicker than XP; and, everything worked immediately. While I know that there are the ocassionaly hardware problems with Ubuntu's installation, It was the smoothest thing to Windows that I have experienced. Since then I have found that the Ubuntu and Debian communities provide tons of useful information for just about everything I want to do; and, there are Ubuntu or Debian packages available for just about every piece of software I have wanted
I have liked it so much that I have just removed the Windows partition from my system since I had not used it in over 6 months. I've had great experiences with developers from Gnome, Ubuntu and other projects when I have reported a bug; most of them getting fixed within a few weeks, not an entire release cycle like commercial software.
I just got a DVD burner for my PC; and, somewhat to my surprise, burning a DVD was easier that the process was for the first time in Windows. My MP3 player, a Creative ZEN something, works flawlessly; as does the scanner, the digital camera, etc. Since I manage Windows servers and clients, the fact that Gnome's RDP client works flawlessly is a blessing.
I thoroughly enjoy Ubuntu; and, I give live CDs to anyone who is interested in experimenting; but, as many other posters have mentioned, trying a few different major distros is a very good practice. I would not buy a car without test driving several makes and models and I only spent ~1.75 hours a day in my car. I spend more time with my OS; so, that choice impacts my daily experience even more.
I wish you luck finding the perfect distro for you.
One of my "bosses", who is responsible for security control of our Active Directory domain was utterly mystified that I could log into a dozen or so PCs in 4 different offices and change the Domain that they belong to without leaving my chair. She had never heard of VNC let alone FastPush. Her expectation was that I would physically walk to each desk and do this in turn.
With that level of comprehension and that level of responsibility, it would be SO easy to scatter red herrings throughout the enterprise. Without details on how the evidence was gathered and the skill of the investigators, I have little faith in its reliability.
So, what kind of "special" cores would be useful for:
- gaming
- development
- database servers
- terminal servers
- streaming media servers
- artificial intelligence?
What I mean is, what kind of specialized computations would significantly improve the performance of tasks that are efficiently implemented in a co-processor?
Specifically, is there asterisk support for these? I did a quick search and could not determine that either way. I'd love something more affordable than the Digium boards; but, I prefer to use asterisk for VIOP.
Not that this means a damn thing globally; but, when I moderate, I tend to mod up comments that have taught me something I did not know that was relevant to the article. My favorite posters are those who are experts in their field, whatever that may be, who are willing to translate their understanding of a topic into something that makes it understandable by me, the reader.
Anyway, thanks to all those who have taken the time to share what they know with the rest of us.
A couple of days ago I got an end-user support call. "There is a wierd toolbaron my internet and I can use my favorites." My reaction, and that of the others in IT was, "Oh, joy, spyware." So I went over to the desktop in question, removed CoolWebSearch, WebShots and some other malware. I asked the user if the system was as she expected. "No, there is that wierd toolbar there pointing to the IE7 tab bar and my favorites are gone."
Doh!
Windows updates had installed IE7 and that bothered her more than the variety of spyware that had been on her system. Now, one of my projects is getting SUS and SMS up and running so that IT can control updates rather than local machines; but, it is interesting that a supposedly improved browser causes more problems that malware.
Take a hypothetical USB flash drive. It's a really small item, and, it could be put in a very small box of about 4"x1.5"x1". Normally, though, it is packed in a bubble that takes up 5"x4"x1". That's a 333% increase. Now consider that this USB key probably spent a fair amount of time in diesel powered vehicles. First on the ride from the OEM in some Chinese city to the port. Then in a ship to the Port of Oakland. Then in another truck to your local Best Buy.
Since the shipping container that it is traveling in likely has about 540 sqft of capacity, it can hold about 930000 boxed USB keys or about 280195 bubble-packed keys. If the container needed 500 gallons of diesel to move from source to destination, the boxes required 2 milliliters of fuel each while the bubble-packs take over 6.7 mL each.
4.7 mL of diesel does not sound like much, but, if you think about all the shit a store like Best Buy, Walmart, NewEgg, etc. sells in a year, that's enough to run my heater 'til global warming makes heating unnecessary. Add the transport fuel to the fuel needed to initially create the plastic bubble polymers and the fuel needed to run the bubble creation machinery; and, you have a lot of oil.
I'm sure that "shrink" is a cost; but, packaging itself is not free and neither is transportation or waste disposal. My uninformed guess is that elimination of excess packaging beyond the minimum necessary to protect the product would do more to reduce the US's oil consumption than the elimination of SUVs.
Does anyone know how/if this was related to their GroupWise product?
GroupWise is far from Exchange in nearly every possible measure, and the client makes one long for Outlook,
Back in the late 90's I really pushed GroupWise over Exchange; mostly losing that debate as the NT/Win2K tide swept through the corporate IT landscape. Now, after administering Exchange in dozens of installations, from my 2 user home LAN to ~1000 PC organizations, I'm back in an environment where the transition from GroupWise to Exchange has yet to happen.
Man, does GroupWise SUCK!!! I'm there working on application deployment automation, so my experience with the collaboration server is strictly as an end user. From that perspective, it is clear that Novell made very limited, incremental progress in the last 5-6 years while MS turned Exchange into a dramatically better product.
I'd hope that open sourcing GroupWise would allow the development of GW to accelerate from the efforts of the community to the point where it was really competitive on a user-facing and administrator-facing basis with Exchange. I am admittedly new to *NIX, coming into IT via NetWare and then moving to Windows, but, I'd really like to see a open-source variant of the Windows Small Business Server (SBS) package. Something for really small businesses that gives groupware, authentication, file/print, web/application, database servers in a "Just install from the CD" way.
I know that there are OSS alternatives for all of these things. I'm pretty sure that the Debian or Ubuntu "server" installs include all of these things already. But, the one thing that SBS offers is a generally useable default configuration out of the box. 8 hours setup work and a small business network can be running with all of the SBS features working.
I'd like to see an equally simple alternetive in the OSS world. If anyone knows of one, I'll test it out as soon as I can.
Re:Can someone explain this to me?
on
Leopard Vs. Vista
·
· Score: 1
Remember, it wasn't until Apple showed the industry that case design mattered that PCs had anything other than big beige boxes with noisy fans.
Seems that there were some pretty damn cool looking SGI boxes before Apple tripped onto the VW Beetle bandwagon.
Re:Because there is no enforcement.
on
Leopard Vs. Vista
·
· Score: 2, Insightful
I am not an OEM accountant; but, I strongly suspect that support costs are a major reason for single platform (Windows) offerings from the likes of Dell and HP. Offering a single linux distro (i.e. Fedora, Suse, Ubuntu) would mean that they would have to employ, or contract, two separate end-user support teams. Offering a variety of OS's, (i.e. Windows + Fedora + Ubuntu + Solaris x86) would be an absolute nightmare.
I've been on the phone with OEM hardware support tech who can't co beyond the "Start Button, Settings,..." script. Tell them that you are getting Event XXXX in the Event Log; and, they think you are speaking Greek. If you had people calling saying either, "The Internet keeps crashing when I try to play poker," or "It says 'Kernel panic and won't do anything'" those poor schmucks would lose their minds.
So, you would have to parse out the Windows and Linux calls; and, try to deal with the people who do not know what they are using even though there are big stickers all over the damn box. My guess is that that would more than double the support costs; and, over the support period of the PC, that would probably be more than even a "full OEM" license for Windows.
And, I know that there are tons of free resources from Red Hat, Canonical, Novell, and the vast developer and user communities; but, for people who think that their OS is Microsoft Office and that the Internet crashes; they might as well be printed man pages on Mars. It seems like the default American response is to pick up the phone and yell until you get what you want.
Around the time I had a programmable TI calculator with a BASIC interpreter and 4kB of RAM, I had a "word processor" for typing papers for school. It was basically an electric typewriter body, complete with replaceable daisy wheel for "fonts". The keyboard input went into a buffer that was viewable through a tiny (~2 rows x 40 chars) where you could make post-typing edits like soft word processors. Once you were finished typing and editing, you could either print to the daisy wheel or save to floppy. If I recall correctly, the floppies contained marked-up text, so, if you could use them on different machines from the same OEM.
Much nicer than a basic typewriter for someone who is not an accurate typist; but, not a computer by any means.
As for your Linksys wireless card, they have broadcom chipsets which are supported for the most part. You'll need the firmware for the card though. Here's a doc for how to install it on Ubuntu. Whatever distribution you use you'll need to use the fwcutter utility and the firmware to get a broadcom card working.
With Windows, I pop in the CD from the NIC's (or PC's) box; click OK a couple of times; and, I'm done. The whole process takes a perhaps 2-3 minutes and very little effort beyond noticing the "INSERT THIS CD" in 144 point type.
Finding another PC with a working network connection, downloading fwcutter, printing the how-to, burning the drivers and fwcutter to CD (after finding a blank CD in the closet), going back to the Linux box, following the how-to...
It seems utterly silly to me that anyone would propose this a comparable to Windows. I'm willing to do this sort of thing because I like to. I'm building a weather station out of Dallas 1Wire sensors and an serial interface box with a customized version of thermd. For fun.
But, my friend, who wrote code for the original digital telecom switches in Fortran on punch cards, is hesitant to install Ubuntu because he is comfortable with Windows, MS Office, etc. and is focusing his energy on remodeling his house, not learning a new OS and apps. If he is not willing to give it a go; what "normal user" is?
And, why do some feel compelled to suggest that because a task is possible in the FOSS world, that it is automatically comparable to the M$ world. Sorry for the rant; I'm up too late.
Frankly, I'd like to see the tags go away. I've yet to see them used in a way that might be useful for semantic searching or anything else. Just my 2 cents...
I'd like to thank the parent for this comment. My experience as tech support for my graphic designer wife is that the publishing industry is a total pain to deal with even when you are using the same OS and app suite. One printer wants PDF done this way, one wants Quark files, one provides accurate color profiles, another makes me do "guestimate profiles by comparing prints to the original file and creating kinda-right correction curves by hand.
I've been unable to figure out how to make our colorimeter work to generate a profile for my Ubuntu system; and, Scribus feels like going back to PageMaker on Mac OS 6 in the late 80s. I use the GIMP for messing with my personal digital pix; but, for anything going to press, Creative Suite / Xpress are the only viable solution I have discovered.
Any bets on how long until Adobe ports CS to Linux?
Slashdot Mirror
My perception is that there is a optimal balance between choice and usefulness. If we were living in a world where the only languages were Assembler and BASIC programming would be hugely difficult and software would be suck because the tools were so mismatched to the application. On the other extreme, if there was a language created for every project, programming would be hugely difficult and software would suck because not only would the application have to be written, but the language as well.
Somewhere in the middle is the perfect balance between endless variety of languages and extreme limitations. My feeling is that if some of the effort by the very intelligent and skilled developers who are creating new languages were instead channeled into improving existing languages, the limitations of current tools would be greatly reduced and the experience of all those who know existing languages would be relevant to solving problems. If a developer has to devote time and effort learning a new language, basic syntax, code conventions, optimal design patterns, etc., then that developer is losing time actually writing useful code.
I have written "Hello World" or some generic "Music Store" type app learning:
- BASIC
- Pascal
- HyperCard
- Fortran
- C
- C++
- Java
- Perl
- PHP
- Python
- DOS/CMD Batchfiles
- Shell Scripting
- Kixtart
- Ruby
- Visual Basic
- C#
- MSH/Monad/Microsoft's new shell
- Lisp
Now, I am not a programmer by profession. I've taken courses and I like to write little programs for personal use; so, my knowledge of these languages is minimal. But, I have had to learn what I know so that I could use existing code or frameworks to accomplish what I wanted. The same requirements for redundant learning applies to those who write code on a professional level. If a highly skilled C++ programmer wants to contribute to a project written in MONO/C#, there is a learning curve. If he then wants to work on a project in Python, another learning curve.I would be willing to bet that some projects miss out on a lot of contributed code because people who are willing and have the algorithmic skills to assist are simply unwilling to take the time to learn whatever language is used.
Sorry it this was less than readable. It is late and I am tired and hungry. Off to the kitchen...
I started my career in IT installing a NetWare server in a trucking company that had previously used Wyse terminals connected to an IBM big-iron box in the early 90s. During most of the 90s I worked on Novell NetWare networks in companies like Westinghouse, Bank of America, GTE, Marriott Intl., etc. It was only when Windows 2000 came out that the customers I worked with started leaving the NetWare fold. Even now, I just finished a gig with a municipality helping them finish their transition from NetWare to 2003 Server. And, in 2007, they still have not migrated away from GroupWise.
So, though NetWare is a legacy product, Novell is very well know, and, in many cases liked, by corporate IT managers. I have very fond memories of using tools like ZenWorks to roll out hundreds of Windows 98 clients with little hands-on effort while RIS and SMS were just a gleam in Mr. Balmer's eye. Though I have no data to support it, I would be willing to bet that Novell has a lot more brand recognition than Red Hat in the enterprise market.
I remember those kits that made Fieros look like Ferraris back in the 80s. I do not think that Ferrari's reputation suffered a bit; though, I do not think much of the schmucks that bought the kits. As long as no one is selling the "skinned" phones as if they were legit iPhones, I'm not sure where the harm to Apple exists.
Wow, you puppy can type. I've tried, but mine gets too distracted trying to lick the cursor off the screen to get beyond the home keys.
I was wondering if any legal history scholars happen to be slashdot regulars who could explain the parallels (or lack thereof) between Alcohol Prohibition and the current efforts to restrict digital content. It seems that there should be some relationship there, though I have nowhere near the education to analyze it myself. Thanks.
I just went to RaLink's site; and, I'd buy a RT2800P based PCI NIC to test. The only problem is, I cannot find who retails these devices. It is hard to support these companies when it seems that noone will link to actual purchaseable products. Even google fails to come up with a vendor when searching for RT2800P.
When I go to some of the non-cooperative OEM sites, they provide links to retailers. And, I can find tons of stuff on newegg.com with the closed firmware hardware; but, nothing with open firmware.
Can someone point me to a US retailer that sells the products that this article suggests, please?
For what it's worth, this is my transition experience... I have been working as an IT professional first with NetWare on the server and DOS/Win9x clients and for the last few years in totally Windows environments. My current project is automating software distribution and maintenance for a municipal government with tools like SMS, RIS, etc. In the 99-00 timeframe, I had downloaded and dual-booted RedHat's distro for a few months. After the novelty wore off, I ended up never using Linux, so, I eventually gave the experiment up. About 8-9 months ago, I decided to give it another go. My initial try was Gentoo, because it seemed sensible to use a source based distro for open-source software. Well, it was cool in a "taking apart the old mechanical clock" sort of way; but, it took forever to download and compile everything. Now, I know that if I had been more careful about my approach it might have gone somewhat quicker; but, a multi-day OS install while googling on another PC was not what I wanted. So, I tried Fedora, Slackware, SuSE, Debian and Ubuntu. I really like both Debian and Ubuntu (a Debian derivative) because of the apt package management tools. I had difficulty with Debian because I do not have a wired ethernet connection to my desk; and, having to pick the PC, monitor and keyboard up and go to the router was not convenient, and, Debian does not have the non-free wireless drivers I need as part of their base installer CD. Ubuntu, on the other hand, installed quicker than XP; and, everything worked immediately. While I know that there are the ocassionaly hardware problems with Ubuntu's installation, It was the smoothest thing to Windows that I have experienced. Since then I have found that the Ubuntu and Debian communities provide tons of useful information for just about everything I want to do; and, there are Ubuntu or Debian packages available for just about every piece of software I have wanted I have liked it so much that I have just removed the Windows partition from my system since I had not used it in over 6 months. I've had great experiences with developers from Gnome, Ubuntu and other projects when I have reported a bug; most of them getting fixed within a few weeks, not an entire release cycle like commercial software. I just got a DVD burner for my PC; and, somewhat to my surprise, burning a DVD was easier that the process was for the first time in Windows. My MP3 player, a Creative ZEN something, works flawlessly; as does the scanner, the digital camera, etc. Since I manage Windows servers and clients, the fact that Gnome's RDP client works flawlessly is a blessing. I thoroughly enjoy Ubuntu; and, I give live CDs to anyone who is interested in experimenting; but, as many other posters have mentioned, trying a few different major distros is a very good practice. I would not buy a car without test driving several makes and models and I only spent ~1.75 hours a day in my car. I spend more time with my OS; so, that choice impacts my daily experience even more. I wish you luck finding the perfect distro for you.
One of my "bosses", who is responsible for security control of our Active Directory domain was utterly mystified that I could log into a dozen or so PCs in 4 different offices and change the Domain that they belong to without leaving my chair. She had never heard of VNC let alone FastPush. Her expectation was that I would physically walk to each desk and do this in turn. With that level of comprehension and that level of responsibility, it would be SO easy to scatter red herrings throughout the enterprise. Without details on how the evidence was gathered and the skill of the investigators, I have little faith in its reliability.
So, what kind of "special" cores would be useful for:
- gaming
- development
- database servers
- terminal servers
- streaming media servers
- artificial intelligence?
What I mean is, what kind of specialized computations would significantly improve the performance of tasks that are efficiently implemented in a co-processor?
I recommend starting with Crawford.
I didn't know you could encode ILUVTITS in 4 digits... Wow.
Specifically, is there asterisk support for these? I did a quick search and could not determine that either way. I'd love something more affordable than the Digium boards; but, I prefer to use asterisk for VIOP.
Not that this means a damn thing globally; but, when I moderate, I tend to mod up comments that have taught me something I did not know that was relevant to the article. My favorite posters are those who are experts in their field, whatever that may be, who are willing to translate their understanding of a topic into something that makes it understandable by me, the reader.
Anyway, thanks to all those who have taken the time to share what they know with the rest of us.
A couple of days ago I got an end-user support call. "There is a wierd toolbaron my internet and I can use my favorites." My reaction, and that of the others in IT was, "Oh, joy, spyware." So I went over to the desktop in question, removed CoolWebSearch, WebShots and some other malware. I asked the user if the system was as she expected. "No, there is that wierd toolbar there pointing to the IE7 tab bar and my favorites are gone."
Doh!
Windows updates had installed IE7 and that bothered her more than the variety of spyware that had been on her system. Now, one of my projects is getting SUS and SMS up and running so that IT can control updates rather than local machines; but, it is interesting that a supposedly improved browser causes more problems that malware.
Take a hypothetical USB flash drive. It's a really small item, and, it could be put in a very small box of about 4"x1.5"x1". Normally, though, it is packed in a bubble that takes up 5"x4"x1". That's a 333% increase. Now consider that this USB key probably spent a fair amount of time in diesel powered vehicles. First on the ride from the OEM in some Chinese city to the port. Then in a ship to the Port of Oakland. Then in another truck to your local Best Buy. Since the shipping container that it is traveling in likely has about 540 sqft of capacity, it can hold about 930000 boxed USB keys or about 280195 bubble-packed keys. If the container needed 500 gallons of diesel to move from source to destination, the boxes required 2 milliliters of fuel each while the bubble-packs take over 6.7 mL each.
4.7 mL of diesel does not sound like much, but, if you think about all the shit a store like Best Buy, Walmart, NewEgg, etc. sells in a year, that's enough to run my heater 'til global warming makes heating unnecessary. Add the transport fuel to the fuel needed to initially create the plastic bubble polymers and the fuel needed to run the bubble creation machinery; and, you have a lot of oil.
I'm sure that "shrink" is a cost; but, packaging itself is not free and neither is transportation or waste disposal. My uninformed guess is that elimination of excess packaging beyond the minimum necessary to protect the product would do more to reduce the US's oil consumption than the elimination of SUVs.
Back in the late 90's I really pushed GroupWise over Exchange; mostly losing that debate as the NT/Win2K tide swept through the corporate IT landscape. Now, after administering Exchange in dozens of installations, from my 2 user home LAN to ~1000 PC organizations, I'm back in an environment where the transition from GroupWise to Exchange has yet to happen.
Man, does GroupWise SUCK!!! I'm there working on application deployment automation, so my experience with the collaboration server is strictly as an end user. From that perspective, it is clear that Novell made very limited, incremental progress in the last 5-6 years while MS turned Exchange into a dramatically better product.
I'd hope that open sourcing GroupWise would allow the development of GW to accelerate from the efforts of the community to the point where it was really competitive on a user-facing and administrator-facing basis with Exchange. I am admittedly new to *NIX, coming into IT via NetWare and then moving to Windows, but, I'd really like to see a open-source variant of the Windows Small Business Server (SBS) package. Something for really small businesses that gives groupware, authentication, file/print, web/application, database servers in a "Just install from the CD" way.
I know that there are OSS alternatives for all of these things. I'm pretty sure that the Debian or Ubuntu "server" installs include all of these things already. But, the one thing that SBS offers is a generally useable default configuration out of the box. 8 hours setup work and a small business network can be running with all of the SBS features working.
I'd like to see an equally simple alternetive in the OSS world. If anyone knows of one, I'll test it out as soon as I can.
Yahoo is Skippy and Google is Adams Natural?
I am not an OEM accountant; but, I strongly suspect that support costs are a major reason for single platform (Windows) offerings from the likes of Dell and HP. Offering a single linux distro (i.e. Fedora, Suse, Ubuntu) would mean that they would have to employ, or contract, two separate end-user support teams. Offering a variety of OS's, (i.e. Windows + Fedora + Ubuntu + Solaris x86) would be an absolute nightmare. I've been on the phone with OEM hardware support tech who can't co beyond the "Start Button, Settings,..." script. Tell them that you are getting Event XXXX in the Event Log; and, they think you are speaking Greek. If you had people calling saying either, "The Internet keeps crashing when I try to play poker," or "It says 'Kernel panic and won't do anything'" those poor schmucks would lose their minds. So, you would have to parse out the Windows and Linux calls; and, try to deal with the people who do not know what they are using even though there are big stickers all over the damn box. My guess is that that would more than double the support costs; and, over the support period of the PC, that would probably be more than even a "full OEM" license for Windows. And, I know that there are tons of free resources from Red Hat, Canonical, Novell, and the vast developer and user communities; but, for people who think that their OS is Microsoft Office and that the Internet crashes; they might as well be printed man pages on Mars. It seems like the default American response is to pick up the phone and yell until you get what you want.
Around the time I had a programmable TI calculator with a BASIC interpreter and 4kB of RAM, I had a "word processor" for typing papers for school. It was basically an electric typewriter body, complete with replaceable daisy wheel for "fonts". The keyboard input went into a buffer that was viewable through a tiny (~2 rows x 40 chars) where you could make post-typing edits like soft word processors. Once you were finished typing and editing, you could either print to the daisy wheel or save to floppy. If I recall correctly, the floppies contained marked-up text, so, if you could use them on different machines from the same OEM. Much nicer than a basic typewriter for someone who is not an accurate typist; but, not a computer by any means.
With Windows, I pop in the CD from the NIC's (or PC's) box; click OK a couple of times; and, I'm done. The whole process takes a perhaps 2-3 minutes and very little effort beyond noticing the "INSERT THIS CD" in 144 point type.
Finding another PC with a working network connection, downloading fwcutter, printing the how-to, burning the drivers and fwcutter to CD (after finding a blank CD in the closet), going back to the Linux box, following the how-to...
It seems utterly silly to me that anyone would propose this a comparable to Windows. I'm willing to do this sort of thing because I like to. I'm building a weather station out of Dallas 1Wire sensors and an serial interface box with a customized version of thermd. For fun.
But, my friend, who wrote code for the original digital telecom switches in Fortran on punch cards, is hesitant to install Ubuntu because he is comfortable with Windows, MS Office, etc. and is focusing his energy on remodeling his house, not learning a new OS and apps. If he is not willing to give it a go; what "normal user" is?
And, why do some feel compelled to suggest that because a task is possible in the FOSS world, that it is automatically comparable to the M$ world. Sorry for the rant; I'm up too late.
Frankly, I'd like to see the tags go away. I've yet to see them used in a way that might be useful for semantic searching or anything else. Just my 2 cents...
I'd like to thank the parent for this comment. My experience as tech support for my graphic designer wife is that the publishing industry is a total pain to deal with even when you are using the same OS and app suite. One printer wants PDF done this way, one wants Quark files, one provides accurate color profiles, another makes me do "guestimate profiles by comparing prints to the original file and creating kinda-right correction curves by hand. I've been unable to figure out how to make our colorimeter work to generate a profile for my Ubuntu system; and, Scribus feels like going back to PageMaker on Mac OS 6 in the late 80s. I use the GIMP for messing with my personal digital pix; but, for anything going to press, Creative Suite / Xpress are the only viable solution I have discovered. Any bets on how long until Adobe ports CS to Linux?
http://www.jcxsoftware.com/jcx/vsphp/home is a PHP plugin for Visual Studio. I used it briefly before switching to ubuntu eclipse and phpeclipse.