I'm a developer with a CS bachelors and about 8 years of post-college work experience. Here's my take on your questions:
1. Over the long haul, you will find many collegues without degrees making as much as you are. The benefit a degree gives you, however, is increased hireability. Your non-degreed compatriots will stand to lose a lot of seniority ($$$) when moving to another company, unless they get "pulled" from another company. Don't expect your university to teach you computers, though. Spend lots of time in the labs and in your dorm room writing your own pet projects and tinkering with Linux/Windows/etc.
2. EE degrees are more specialized, and if you enjoy hardware more than software, go for it. Keep in mind though that EE work can be much more regional, whereas software development jobs can be found in just about any town greater than 40K people.
3. I found college courses to be infinitely better than high school, particularly the "core" English/writing and History courses. In my English 101 course, for instance, we watched and analyzed the movie "Angel Heart". Try that in a a public high school. My History 101 prof always began the day with a (generally dirty) joke of the day. Much more adult atmosphere. These core classes can be very large too (150+), so if you can pull off the exams you can sleep in.
4. Probably not, but the key to scholarships is to get lots of small ones, which is how lots of "fully-paid" students get fully paid: with a dozen small scholarships for "citizenship" or whatnot.
5. I'd definitely go with the degree, but I'd tell myself that I'm going to make squat (25-30K) when I graduate, but that changes over the next few years. At five years exp., you jump to the "senior" status and should look for another job if you're not in the 60-80K+ range.
I've been meaning to write a script that will compare the current kernel version with the last kernel booted (probably use 'uname -a >/etc/lastkernel at the end of rc.local), and if it differs, run the rpmbuild --rebuild scripts. You could put this script in rc.local and not have to worry about rebuilding the NVIDIA drivers when you update your kernel.
I'll do this in my copious free time, but maybe I'll seed the idea here and someone else will do it first.
Re:A fork would be *bad*
on
XFree86 Politics
·
· Score: 2, Insightful
A forked tree would basically kill X as a _standard_ platform. If I write an X app, do I write it for XFree86 or XFree-KiethP?
Considering that I can run every Linux app imaginable using a completely non-XFree86 X11 server on Windows (XWin-Pro, Xceed), I'd have to disagree with you there. I'm sure the KDE team didn't "write KDE for" XWin-Pro, but it works fine. Also, there are non-XFree86 X servers for Linux that work great, one of which I can't remember the name (it's a commercial server).
X11 is the standard. XFree is just an implementation of the the standard. Similar to HTTP is the standard, but Apache and IIS are implementations.
...DirectX was the first graphics programming library to really make it easy to create complex graphics...
Sorry, but I've got to heavily disagree here. OpenGL is one of the nicest API's I've worked with; I'd place it as a model of how to build a good C API. Namely, the system initializes in a known good state, and the programmer changes that state to suit. It's a much easier system to get started with, and it sounds a lot like good OO programming, in fact (create an object in a known good state, then alter properties to suit).
DirectX, and Direct3D in particular, have that annoying Microsoft API habit of requiring umpteen structs (often with embedded structs) to be filled and/or initialized, and then passed in to get anything done. I've heard the latest incarnations have begun to move away from that, but I don't go to MSDN much anymore.
Interesting, I had no idea people were rendering 3D remotely for realtime display:)
Well, I used to have a cool job like that, but I was greedy (stupid) and took more money to do real work (boring).:-)
SGI had to make GL network transparent because at the time (early 90's) only large machines could do the things they needed to do, so you had to hang xterms off a big box.
Even so, I think that OpenGL still does not need a tie to something so complex as an X server, as polygons have nothing to do with TCP/IP. In the case of Linux, there could be an OpenGL driver capable of rendering to video surfaces (DirectFB full/windowed) or to system memory for sending over the network. Assuming you're using an X server that rides on top of DirectFB, all the bases are covered, right?
I'm going into waters that I know little about, but I thought that X was already optimized for local displays in much the way you're describing... isn't that what DRI (Direct Rendering Interface) and other modules do? I know that's why I can get many frames per second with NVidia drivers under xfree86.
Anyway, xfree86 implementation issues aside, don't discount good remote display capabilities... they can be extremely useful in home and in production environments. Also, projects like The Linux Terminal Server Project allow many nice systems to be built very cheaply for schools, home, and business using X11. The best part is that you have all this stuff built-in and ready in your Linux system... get an old Pentium or PII ($50) with a decent PCI graphics card ($50-$75) and 17" monitor ($120) and hang it off your big box on the network... for everything but page-flip games you won't be able to tell you're running remotely. Nice price for a second PC that requires practically no administration, and can even be floppy based (ltsp).
There are too many incompatible toolkits And just how is producing yet another even less compatible programmer's API going to solve this? You're better off trying to invent and push the One True Toolkit, because at least then the transition can be gradual.
I'll throw my $0.02 in with your rant, as well. People who claim that Windows apps all use the same widgets have obviously not used many Windows apps... many of them use their own custom widget sets as well.
Here's a test... load Office XP [shudder]. Note that regardless of the Windows settings, Office uses its own XP-style menus. How could this be if it wasn't implementing its own menus? Next, load Windows XP [double shudder]. Note that the drag behaviour of the standard combo box has changed... you can't click, hold, drag, and release. You must click, release, then click to select an item using the standard combo box (in say, a control panel app). Now, go to Office XP and open an Office dialog box (Tools | Options). The combo box works in the old way -- you can click-drag-release. It's implementing its own controls, or altering the standard ones.
Why the hell does OpenGL, something you're never going to use remotely, require XFree86? Because everyone uses X!
Because OpenGL is meant for more than games. What if you need to render and manipulate a large quantity (huge...gigs of data) of remote sensing data in a 3d environment in realtime? In other words, you need a big honking machine that you can't get at BestBuy and put on your desk. You need a large cluster or SGI Reality machine that's being cared for in its own climate-controlled pen... it handles the massive data crunching while you use your NVidia (or other) desktop to remotely render the OpenGL commands it sends over the network.
It's really amazing how efficient this works... you should try it. I've used it in the instance described above (except the X-term was an SGI box), and I've also used it to get super-smooth framerates on the gl screensavers running on a PII-266 but displayed remotely on an Athlon/GeForce box over a 100mbit network. Things like glearth will run at 60+fps and only use about 60KB/sec of network bandwidth.
The important thing is that X is built for the long haul, and so it must include remote operation... the desktop configuration may not be the ideal solution in the future, and you don't want to limit yourself to the "one user, sitting next to the box" configuration over the decades (just ask Microsoft, who is trying desperately to turn single-user Windows into multi-user Unix).
It also needs a much better way of cut&paste for exmple you cant highlite one piece of text and then select copy then highlite another and select paste, that whole middle click paste thing sucks ass.
Actually, that method of cut-and-paste is a "bonus" on top of the clipboard functionality provided by the applications (GTK and Qt use essentially the same methods, which I believe is based on the XDND protocol). You should really only use the middle mouse button for simple text and for swapping text between terminals and such. For anything more complex (such as graphics, spreadsheet cells, etc) use CTRL-X/C/V or the edit menus.
As a power user, I love the convenience of the middle-button clipboard, and I get frustrated at apps that don't honor it (kdevelop, for one).
It also needs a way to change res on the fly but I hear thats going to be in 4.3.
Honestly, I haven't missed this feature, but early on I set up a user called "gamer" running a small window manager at 800x600 and I log into that user when I want to play games. I don't know why else I'd want to change the res on-the-fly, but as you said, they're working on it so I'll take it.
kde adds much of these things but they only work within kde not to mention kde is too much of a mem hog on slow systems
You may not realize that you can run all the KDE apps you want from any window manager. I run kdevelop, kmail, and konqueror all the time, but I don't run the KDE desktop (too slow for me). I use WindowMaker. I also use Gnome apps alongside the KDE apps... these desktop systems are not mutually exclusive! You may realize this, but a lot of new users coming from Windows don't realize that you can run Gnome apps within KDE and vice-versa because the "desktop environments" are pretty much veneers over the standard X protocol and are not incompatible.
Depends on the end-user. I was given a new PC at work with XP, and it was useless for me (as a developer) until I installed Cygwin.
Few times have I felt as patronized as when I hit CTRL-F to search for a file and a stupid dog appeared to `help' me find it. After much digging around the `easy` web-style options I discovered how to actually search without the dog, and how to search (grep) for source files with certain text. Then I discovered that I can't search within files unless some plugin thing has been installed in the system. Did Find Files tell me this? No, after several minutes of no hits, I finally entered something like "TObject" (I was searching a tree of Delphi source code), which returned no hits (TObject should have returned just about every.pas file in the tree). NT4, Win2K, heck, even Win95 would have done this easily.
Shall I go on with the find files? Once I've managed to find the files, I have to now operate three items to get back to my original file manager view... First, close the search (I used to just hit escape). Next, click the "Back" button or backspace to make the search results disappear (don't hit the "Up" button since that won't do what you want). Finally, since it conveniently changed your view to "Details", you now have to change back to "List" or "Icon" view. All this to accomplish:
#>find . -name "*.pas"
or
#>fgrep -r TObject
If I want to preview a 3KB sound file, I have to painfully wait until that hideous MediaPlayer is loaded. I'm sure MPlayer32.exe is somewhere, and I can reconfigure the system, but I thought I was supposed to be getting an advantage from using this system over Linux? What, I've paid good money, lost my freedoms and I have to tinker with my system to make it work?
There are many other such productivity killing items that make this system little more than a toy OS good only for launching applications, but I'm home now and don't want to think about that ridiculous system at work. Basically, in my mind, Microsoft has forsaken me and my daily productivity (and happiness) so that some mythical grandma can "check out that internet". The scary thing is that I have XP Pro... I'd hate to see how squishy the Home edition is. Frankly, I don't care how easy it is for some home user to download pr0n and mp3's... I have work to do! Microsoft will get what they deserve over the next decade.
Now, let's look back in the day. WP5.1, DOS 5.0, and Lotus 123. WYSIWYG+123 was not much better. Those applications required training and complicated Function Key cards above the KB.
This reminded me of when I did admin work in college for a law firm. There were about 20-30 clerical workers, all running WordPerfect 5.2 for DOS. In a year of work, I only once had to help one of them with the software... usually I just had to help them with their Inkjet jams (incidentally, I don't ever recall the software crashing).
However, in times since then, I've had to help dozens of people accomplish fairly simple tasks with Word. Were the clerical workers that well trained? Were the wierd CTRL-SHIFT-F7 controls actually easier once learned? Or, perhaps we no longer have pure "clerical workers" like before, and each person is expected to (poorly, it seems) handle their own dictation and other clerical work. I have no idea. I'm a lot more productive with GUI word processors over the old WordPerfect, but then I only use a word processer a handful of times per year.
// FIXME: We have no idea what this does, // but we're afraid to touch it. It caused // an infinite loop in the Eden testing lab. // See workaround below. -Adam 1.0 team
Brain::hiccup() { while (1) { // old code. don't touch. memcpy(GLOTTIS, 0xff); sleep(2000 * (random() + 0.5));
// FIX added to work around infinite loop if (fearLevel() > 0.7) break; } }
...are price, and battery life. Neither of which were addressed on this web page. If it's $200 and lasts three hours on a battery charge, then we've probably got another Lynx. (is that what is was called? There were a couple of those deluxe handhelds out in the early 90's)
The GameBoy has survived because it's cheap and the battery life lasts forever (well, an impressively long time, at least). That, and the Mario/Zelda/Pokemon franchises.
This is a pretty big name distro, at least in the eyes of Joe Buying Software Off The Shelf.
I never like to see a company with a decent product go under, but frankly, Joe User and Aunt Tillie will just have to wait about 8-10 years to get Linux on their home PC, although they will likely be surrounded by Linux in embedded units. And that's OK: that's not where the real opportunity for Linux lies. Large businesses are the real consumers of Linux, and they weren't even looking at Mandrake (hence the bankruptcy).
Joe User will buy a home PC to match his work system, not the other way around. Make sure he has a Linux box at work and in a few years he'll want one at home to stay compatible in case he "has to take some work home one day". The Quickbooks-oriented small business market is as tough to crack as the consumer market.
It doesn't matter anyway; while the naysayers say "see! I told you so!" and the cheerleaders wring their hands and wonder if the sky is falling, Free/Open Source Software will continue to march along at its own pace, blissfully unaware of the uproar surrounding it.
Not likely....a Sun/Oracle/Borland combination would be a viable IBM competitor, but not a Microsoft killer. The sheer complexity and mass of the resulting organization would sidetrack it for years, while IBM and Microsoft continue on their merry ways.
You are correct -- I meant to specify "in the enterprise market". The new conglomerate wouldn't know beans about the home/SOHO/consumer realms that Microsoft holds.
Although I wish Borland would stay a scrappy independent forever, I always thought that Sun or Oracle would be a good parent for Borland.
Actually, if you could get past the CEO egos, a combined Sun/Oracle/Borland could be an instant IBM and Microsoft "killer". They would have the hardware from desktop on up that could be supplemented with x86 hardware, the enterprise backend software (J2EE, Oracle, etc), some of the best development tools around (Delpi/Kylix for Linux/Solaris, JBuilder, CBuilder, etc), and an Office suite to boot.
Their corporate cultures seem to be compatible, from what I know of them (not much, directly, but based on years of reading). I don't see anything compatible between Microsoft and Borland, however.
You must mean ".NET My Services" because.NET as a technology has been rolled out and has been receiving wide acceptance.
I'm guessing he was talking about marketing to the IT Pros/PHBs, many of whom are still scratching their heads about what to do about it and wondering where the install disk is that will make their software run on Web Services. Others don't want to use.Net because they'll have to use Passport. The rest are now saying "Whatever happened to.Net?" (I was actually asked this by a fairly tech-savvy MBA recently). Yes, I know all these are untrue, but most non-techies don't understand this real purpose of.Net or what it is supposed to be (a product? a new version of Windows? something to do with the Web?)
The truth is, they should have pitched it as a new API replacing Win32. Not as sexy, but much closer to the truth.
M$ has a lot of problems, but this.NET stuff is cool and people should take notice. Even the evil empire can raise the bar. And competition helps us all in the end. Lower those prices SUN and Oracle!!
Yes, and they (MS) are creating their own monster by focusing on price. The more they train the market to look at "lower cost" vs. Oracle and Sun, the more legitimacy they give to Linux, MySQL, PostgreSQL, and PHP vs. Microsoft. Bring it on, I say.
On the other hand, I seriously doubt the 2K vs. 14K lines of code is legitimate. If it is, there is a lot of opaque, drag-n-drop, hope-it-just-works development in the smaller one. A size ratio like that might happen when comparing against maybe C and stdlib for web apps; then again, I think even using printf's and socket calls one could get closer than that.
This reminds me of when I worked as a lab assistant in college. In the two years I worked there, my assistance to users devolved in the following steps:
At first, I would explain the concepts to the user to make sure they understood not just the how, but the why.
Over time, this eventually gave way to just telling them "click this button" or "hit this key".
By the end of my tenure, I would just walk up, slide the keyboard in front of me, type whatever was needed, and then slide it back.
The sad thing is that for most of the patrons there was no difference in the three methods. They simply didn't care to know more than the minimum to get the current task completed. Never mind that you might want to know how to do future tasks better. Some were clueful, of course, and they can usually be detected from 50 paces; I still explained the how and why to them. But the others....
Slashdot Mirror
I'm a developer with a CS bachelors and about 8 years of post-college work experience. Here's my take on your questions:
1. Over the long haul, you will find many collegues without degrees making as much as you are. The benefit a degree gives you, however, is increased hireability. Your non-degreed compatriots will stand to lose a lot of seniority ($$$) when moving to another company, unless they get "pulled" from another company. Don't expect your university to teach you computers, though. Spend lots of time in the labs and in your dorm room writing your own pet projects and tinkering with Linux/Windows/etc.
2. EE degrees are more specialized, and if you enjoy hardware more than software, go for it. Keep in mind though that EE work can be much more regional, whereas software development jobs can be found in just about any town greater than 40K people.
3. I found college courses to be infinitely better than high school, particularly the "core" English/writing and History courses. In my English 101 course, for instance, we watched and analyzed the movie "Angel Heart". Try that in a a public high school. My History 101 prof always began the day with a (generally dirty) joke of the day. Much more adult atmosphere. These core classes can be very large too (150+), so if you can pull off the exams you can sleep in.
4. Probably not, but the key to scholarships is to get lots of small ones, which is how lots of "fully-paid" students get fully paid: with a dozen small scholarships for "citizenship" or whatnot.
5. I'd definitely go with the degree, but I'd tell myself that I'm going to make squat (25-30K) when I graduate, but that changes over the next few years. At five years exp., you jump to the "senior" status and should look for another job if you're not in the 60-80K+ range.
6. Nope, sorry.
I've been meaning to write a script that will compare the current kernel version with the last kernel booted (probably use 'uname -a > /etc/lastkernel at the end of rc.local), and if it differs, run the rpmbuild --rebuild scripts. You could put this script in rc.local and not have to worry about rebuilding the NVIDIA drivers when you update your kernel.
I'll do this in my copious free time, but maybe I'll seed the idea here and someone else will do it first.
A forked tree would basically kill X as a _standard_ platform. If I write an X app, do I write it for XFree86 or XFree-KiethP?
Considering that I can run every Linux app imaginable using a completely non-XFree86 X11 server on Windows (XWin-Pro, Xceed), I'd have to disagree with you there. I'm sure the KDE team didn't "write KDE for" XWin-Pro, but it works fine. Also, there are non-XFree86 X servers for Linux that work great, one of which I can't remember the name (it's a commercial server).
X11 is the standard. XFree is just an implementation of the the standard. Similar to HTTP is the standard, but Apache and IIS are implementations.
over and over and over... guys, there is no command.com on my system. Give it a rest!
...DirectX was the first graphics programming library to really make it easy to create complex graphics...
Sorry, but I've got to heavily disagree here. OpenGL is one of the nicest API's I've worked with; I'd place it as a model of how to build a good C API. Namely, the system initializes in a known good state, and the programmer changes that state to suit. It's a much easier system to get started with, and it sounds a lot like good OO programming, in fact (create an object in a known good state, then alter properties to suit).
DirectX, and Direct3D in particular, have that annoying Microsoft API habit of requiring umpteen structs (often with embedded structs) to be filled and/or initialized, and then passed in to get anything done. I've heard the latest incarnations have begun to move away from that, but I don't go to MSDN much anymore.
Everything game-wise is written in DirectX.
:)
Yeah, except for those obscure titles from that unknown company... what's their name again? I think they're called Id?
Interesting, I had no idea people were rendering 3D remotely for realtime display :)
:-)
Well, I used to have a cool job like that, but I was greedy (stupid) and took more money to do real work (boring).
SGI had to make GL network transparent because at the time (early 90's) only large machines could do the things they needed to do, so you had to hang xterms off a big box.
Even so, I think that OpenGL still does not need a tie to something so complex as an X server, as polygons have nothing to do with TCP/IP. In the case of Linux, there could be an OpenGL driver capable of rendering to video surfaces (DirectFB full/windowed) or to system memory for sending over the network. Assuming you're using an X server that rides on top of DirectFB, all the bases are covered, right?
I'm going into waters that I know little about, but I thought that X was already optimized for local displays in much the way you're describing... isn't that what DRI (Direct Rendering Interface) and other modules do? I know that's why I can get many frames per second with NVidia drivers under xfree86.
Anyway, xfree86 implementation issues aside, don't discount good remote display capabilities... they can be extremely useful in home and in production environments. Also, projects like The Linux Terminal Server Project allow many nice systems to be built very cheaply for schools, home, and business using X11. The best part is that you have all this stuff built-in and ready in your Linux system... get an old Pentium or PII ($50) with a decent PCI graphics card ($50-$75) and 17" monitor ($120) and hang it off your big box on the network... for everything but page-flip games you won't be able to tell you're running remotely. Nice price for a second PC that requires practically no administration, and can even be floppy based (ltsp).
There are too many incompatible toolkits And just how is producing yet another even less compatible programmer's API going to solve this? You're better off trying to invent and push the One True Toolkit, because at least then the transition can be gradual.
I'll throw my $0.02 in with your rant, as well. People who claim that Windows apps all use the same widgets have obviously not used many Windows apps... many of them use their own custom widget sets as well.
Here's a test... load Office XP [shudder]. Note that regardless of the Windows settings, Office uses its own XP-style menus. How could this be if it wasn't implementing its own menus? Next, load Windows XP [double shudder]. Note that the drag behaviour of the standard combo box has changed... you can't click, hold, drag, and release. You must click, release, then click to select an item using the standard combo box (in say, a control panel app). Now, go to Office XP and open an Office dialog box (Tools | Options). The combo box works in the old way -- you can click-drag-release. It's implementing its own controls, or altering the standard ones.
Standard toolkits.... pfft.
Why the hell does OpenGL, something you're never going to use remotely, require XFree86? Because everyone uses X!
Because OpenGL is meant for more than games. What if you need to render and manipulate a large quantity (huge...gigs of data) of remote sensing data in a 3d environment in realtime? In other words, you need a big honking machine that you can't get at BestBuy and put on your desk. You need a large cluster or SGI Reality machine that's being cared for in its own climate-controlled pen... it handles the massive data crunching while you use your NVidia (or other) desktop to remotely render the OpenGL commands it sends over the network.
It's really amazing how efficient this works... you should try it. I've used it in the instance described above (except the X-term was an SGI box), and I've also used it to get super-smooth framerates on the gl screensavers running on a PII-266 but displayed remotely on an Athlon/GeForce box over a 100mbit network. Things like glearth will run at 60+fps and only use about 60KB/sec of network bandwidth.
The important thing is that X is built for the long haul, and so it must include remote operation... the desktop configuration may not be the ideal solution in the future, and you don't want to limit yourself to the "one user, sitting next to the box" configuration over the decades (just ask Microsoft, who is trying desperately to turn single-user Windows into multi-user Unix).
It also needs a much better way of cut&paste for exmple you cant highlite one piece of text and then select copy then highlite another and select paste, that whole middle click paste thing sucks ass.
Actually, that method of cut-and-paste is a "bonus" on top of the clipboard functionality provided by the applications (GTK and Qt use essentially the same methods, which I believe is based on the XDND protocol). You should really only use the middle mouse button for simple text and for swapping text between terminals and such. For anything more complex (such as graphics, spreadsheet cells, etc) use CTRL-X/C/V or the edit menus.
As a power user, I love the convenience of the middle-button clipboard, and I get frustrated at apps that don't honor it (kdevelop, for one).
It also needs a way to change res on the fly but I hear thats going to be in 4.3.
Honestly, I haven't missed this feature, but early on I set up a user called "gamer" running a small window manager at 800x600 and I log into that user when I want to play games. I don't know why else I'd want to change the res on-the-fly, but as you said, they're working on it so I'll take it.
kde adds much of these things but they only work within kde not to mention kde is too much of a mem hog on slow systems
You may not realize that you can run all the KDE apps you want from any window manager. I run kdevelop, kmail, and konqueror all the time, but I don't run the KDE desktop (too slow for me). I use WindowMaker. I also use Gnome apps alongside the KDE apps... these desktop systems are not mutually exclusive! You may realize this, but a lot of new users coming from Windows don't realize that you can run Gnome apps within KDE and vice-versa because the "desktop environments" are pretty much veneers over the standard X protocol and are not incompatible.
Depends on the end-user. I was given a new PC at work with XP, and it was useless for me (as a developer) until I installed Cygwin.
.pas file in the tree). NT4, Win2K, heck, even Win95 would have done this easily.
Few times have I felt as patronized as when I hit CTRL-F to search for a file and a stupid dog appeared to `help' me find it. After much digging around the `easy` web-style options I discovered how to actually search without the dog, and how to search (grep) for source files with certain text. Then I discovered that I can't search within files unless some plugin thing has been installed in the system. Did Find Files tell me this? No, after several minutes of no hits, I finally entered something like "TObject" (I was searching a tree of Delphi source code), which returned no hits (TObject should have returned just about every
Shall I go on with the find files? Once I've managed to find the files, I have to now operate three items to get back to my original file manager view... First, close the search (I used to just hit escape). Next, click the "Back" button or backspace to make the search results disappear (don't hit the "Up" button since that won't do what you want). Finally, since it conveniently changed your view to "Details", you now have to change back to "List" or "Icon" view. All this to accomplish:
#>find . -name "*.pas"
or
#>fgrep -r TObject
If I want to preview a 3KB sound file, I have to painfully wait until that hideous MediaPlayer is loaded. I'm sure MPlayer32.exe is somewhere, and I can reconfigure the system, but I thought I was supposed to be getting an advantage from using this system over Linux? What, I've paid good money, lost my freedoms and I have to tinker with my system to make it work?
There are many other such productivity killing items that make this system little more than a toy OS good only for launching applications, but I'm home now and don't want to think about that ridiculous system at work. Basically, in my mind, Microsoft has forsaken me and my daily productivity (and happiness) so that some mythical grandma can "check out that internet". The scary thing is that I have XP Pro... I'd hate to see how squishy the Home edition is. Frankly, I don't care how easy it is for some home user to download pr0n and mp3's... I have work to do! Microsoft will get what they deserve over the next decade.
Whew. Flame off... it's been a long day.
Now, let's look back in the day. WP5.1, DOS 5.0, and Lotus 123. WYSIWYG+123 was not much better. Those applications required training and complicated Function Key cards above the KB.
This reminded me of when I did admin work in college for a law firm. There were about 20-30 clerical workers, all running WordPerfect 5.2 for DOS. In a year of work, I only once had to help one of them with the software... usually I just had to help them with their Inkjet jams (incidentally, I don't ever recall the software crashing).
However, in times since then, I've had to help dozens of people accomplish fairly simple tasks with Word. Were the clerical workers that well trained? Were the wierd CTRL-SHIFT-F7 controls actually easier once learned? Or, perhaps we no longer have pure "clerical workers" like before, and each person is expected to (poorly, it seems) handle their own dictation and other clerical work. I have no idea. I'm a lot more productive with GUI word processors over the old WordPerfect, but then I only use a word processer a handful of times per year.
Comments?
...
// old code. don't touch.
// FIX added to work around infinite loop
// FIXME: We have no idea what this does,
// but we're afraid to touch it. It caused
// an infinite loop in the Eden testing lab.
// See workaround below. -Adam 1.0 team
Brain::hiccup()
{
while (1)
{
memcpy(GLOTTIS, 0xff);
sleep(2000 * (random() + 0.5));
if (fearLevel() > 0.7)
break;
}
}
...are price, and battery life. Neither of which were addressed on this web page. If it's $200 and lasts three hours on a battery charge, then we've probably got another Lynx. (is that what is was called? There were a couple of those deluxe handhelds out in the early 90's)
The GameBoy has survived because it's cheap and the battery life lasts forever (well, an impressively long time, at least). That, and the Mario/Zelda/Pokemon franchises.
This is a pretty big name distro, at least in the eyes of Joe Buying Software Off The Shelf.
I never like to see a company with a decent product go under, but frankly, Joe User and Aunt Tillie will just have to wait about 8-10 years to get Linux on their home PC, although they will likely be surrounded by Linux in embedded units. And that's OK: that's not where the real opportunity for Linux lies. Large businesses are the real consumers of Linux, and they weren't even looking at Mandrake (hence the bankruptcy).
Joe User will buy a home PC to match his work system, not the other way around. Make sure he has a Linux box at work and in a few years he'll want one at home to stay compatible in case he "has to take some work home one day". The Quickbooks-oriented small business market is as tough to crack as the consumer market.
It doesn't matter anyway; while the naysayers say "see! I told you so!" and the cheerleaders wring their hands and wonder if the sky is falling, Free/Open Source Software will continue to march along at its own pace, blissfully unaware of the uproar surrounding it.
...wearing my tinfoil suit, but who's laughing now?!?
How do we make SCO an example to the world that this is a very bad idea?
I know! From now on, nobody buy their products!
Oh, wait...
Not likely....a Sun/Oracle/Borland combination would be a viable IBM competitor, but not a Microsoft killer. The sheer complexity and mass of the resulting organization would sidetrack it for years, while IBM and Microsoft continue on their merry ways.
You are correct -- I meant to specify "in the enterprise market". The new conglomerate wouldn't know beans about the home/SOHO/consumer realms that Microsoft holds.
Although I wish Borland would stay a scrappy independent forever, I always thought that Sun or Oracle would be a good parent for Borland.
Actually, if you could get past the CEO egos, a combined Sun/Oracle/Borland could be an instant IBM and Microsoft "killer". They would have the hardware from desktop on up that could be supplemented with x86 hardware, the enterprise backend software (J2EE, Oracle, etc), some of the best development tools around (Delpi/Kylix for Linux/Solaris, JBuilder, CBuilder, etc), and an Office suite to boot.
Their corporate cultures seem to be compatible, from what I know of them (not much, directly, but based on years of reading). I don't see anything compatible between Microsoft and Borland, however.
"Asking if a computer can think is like asking if a submarine can swim"
...with a cart that had all four wheels in proper working order.
You must mean ".NET My Services" because .NET as a technology has been rolled out and has been receiving wide acceptance.
.Net because they'll have to use Passport. The rest are now saying "Whatever happened to .Net?" (I was actually asked this by a fairly tech-savvy MBA recently). Yes, I know all these are untrue, but most non-techies don't understand this real purpose of .Net or what it is supposed to be (a product? a new version of Windows? something to do with the Web?)
I'm guessing he was talking about marketing to the IT Pros/PHBs, many of whom are still scratching their heads about what to do about it and wondering where the install disk is that will make their software run on Web Services. Others don't want to use
The truth is, they should have pitched it as a new API replacing Win32. Not as sexy, but much closer to the truth.
M$ has a lot of problems, but this .NET stuff is cool and people should take notice. Even the evil empire can raise the bar. And competition helps us all in the end. Lower those prices SUN and Oracle!!
Yes, and they (MS) are creating their own monster by focusing on price. The more they train the market to look at "lower cost" vs. Oracle and Sun, the more legitimacy they give to Linux, MySQL, PostgreSQL, and PHP vs. Microsoft. Bring it on, I say.
On the other hand, I seriously doubt the 2K vs. 14K lines of code is legitimate. If it is, there is a lot of opaque, drag-n-drop, hope-it-just-works development in the smaller one. A size ratio like that might happen when comparing against maybe C and stdlib for web apps; then again, I think even using printf's and socket calls one could get closer than that.
- At first, I would explain the concepts to the user to make sure they understood not just the how, but the why.
- Over time, this eventually gave way to just telling them "click this button" or "hit this key".
- By the end of my tenure, I would just walk up, slide the keyboard in front of me, type whatever was needed, and then slide it back.
The sad thing is that for most of the patrons there was no difference in the three methods. They simply didn't care to know more than the minimum to get the current task completed. Never mind that you might want to know how to do future tasks better. Some were clueful, of course, and they can usually be detected from 50 paces; I still explained the how and why to them. But the others.......isn't that the planet where they used to build luxury planets for the super-rich?