Actually, Windows 95 (which was delayed over a year) was suprisingly non-buggy, considering it was built on top of the horrific Windows 3.1. It was no NT/OS2/Unix, but it wasn't intended to be either.
Of course, 4 years of accumulated crap has pretty much got Windows 9x back to the stability level of Win3.1. Time to start over again! --
It really depends what you run (or don't run). It's not uncommon for (non-MS) Lotus Domino to be up forever on NT, while (MS) Exchange certainly has it's difficulties.
Despite it's problems, NT4 is the "NetWare 3" of the late 90s LAN - the entrenched market leader. (Before anyone makes any comparison, NetWare 3 wins in many respects, but I remember quite a few abends and configuration nightmares.)
Windows 2000 probably will be a compelling workstation for shops struggling with crappy Win9x, but the server version includes a number of complex and difficult to understand network services aimed at the enterprise market.
This poses a big problem: Larger shops won't want to move to Active Directory (etc) for a while until they've tested the hell out of it. Upgrading is no longer a tactical decision (like NT3.51 to NT4 or NT4 to Linux/Samba), and more of a strategic one which will involve considerable planning and budget and the typical interminable big-IS project BS.
(I don't see much good information from MS about how an Domain to AD transition is really supposed to work. They probably don't know themselves. Furthermore, they've magnified the problem by unifying the Exchange and NOS directories, which makes the prominence of fucking up an order of magnitude greater. Maybe I'm missing something, but I also don't see a way to run Win2000 server in 'Domain-emulation' mode.)
Smaller shops (which by-in-large are running NT with loosely pieced together WINS systems and broadcasts) don't want or need Active Directory, et al, and will have considerable fear and uncertainty over an upgrade. Everything is kinda-sorta working -- why break it.
All of this adds up to a lot of fear and confusion for NT shops. Add in an economic downturn, and IT budget cuts will probably stagante any new project or make cheaper alternatives (ahem) look more appealing.
So, NT5/Win2000 has a big chance at being the "NetWare 4" of the early 00s LAN -- A good portion of the installed base might just skip the upgrade or switch to a less complex and cheaper alternative. --
It's true that people are happy with the workstation aspects of Win2000, but apparently the server stuff (Intellimirror, Active Directory) is still having some issues in a larger environment.
One possiblity is that Win2000 Workstation will be out soon, but the server version will be delayed by a couple months, and the "Datacenter" server (which supposedly supports 32-way SMP), much later. --
Just to clarify what I said - binary font files are distributed under a software licence (GPL, MS EULA, X, etc.), just like any other software. The actual art design of the font can't be copyrighted however. --
Just to make it clear, binary font files still are copyrighted, only the 'design' can not. You can't legally cat YourFont.ttf > MyNewFont.ttf (of course you would need to change some metadata too.)
You can 'copy' someone elses fonts as long as your font drawings are original. For example, you can scan in text and then draw your own outlines around the pictures of the glyphs.
As has been pointed out many times, the *most* important applications to many businesses are the ones that have been developed in house in VB/Delphi/VC++/Access/DBase/Whatever. The likelyhood that corporations would/could port these apps to Linux is pretty low.
In house apps nail the average corporate desktop to Windows, so without something like WINE, you'd probably never see Linux on an average corporate desktop.
(The good thing about most corporate apps is that they're unlikey to use the latest Windows voodoo API, so there's no worry about MS breaking WINE compatiblity.)
There's something to your rant. The other day I popped into comp.os.windows-nt.advocacy (for the first time since the good ol' OS/2 days in '94 + '95), and the on-going Linux/NT was a hellava more intelligent and cordial than a typical Windows NT thread on slashdot. Amazingly, less ads too.
You said it. Lord knows how many NT admins aren't even aware that it's possible to use a DOS boot disk with their machines. (And with Linux Samba servers, btw.) --
Well, Linux might not be competitive with Sun/Solaris server systems yet, but it Linux on commodity x86 systems certainly competes favorably with Sun's (and HP, IBM, etc.) traditional Unix workstation business. SGI has figured this out, which is why they're moving away from Irix/MIPS workstations and towards Linux/x86.
In fact, I would guess that most real Linux desktop deployments (aka real work, and not futzing with HTTP/MP3/IRC) are at the expense of a commercial Unix vendor, and not at the expense of Windows NT. --
ZDNN: So when will you consider that you've 'won'?
Well, when Linux's server market share goes over fifty per cent -- or when Microsoft's stock price crashes, whichever comes first. And again let me interpret that: I'm not saying that Microsoft's stock price crashing is the goal, but when it does, the resistance to doing things right goes away.
We're not taking seats away from Unix, and we're not particularly interested in taking seats away from Unix. Most of the surveys show we're taking seats away from NT -- and I believe Novell has taken a big hit. But we're not killing Solaris. That's OK -- we don't want to kill Solaris.
Since when does Microsoft's stock price have anything to do with the technical goals of Linux? Does RedHat's fat stock increase "resistance" to commercial Unix?
Also, "We're not taking seats away from Unix" !? The hell you're not. How ESR belives that a free unix is not going to take seats away from $$Unix is beyond me.
"We don't want to kill Solaris" - I thought that creating an Open Source clone of Solaris was pretty much the primary design goal of Linux.
Well, to be fair, you had Andresson and others running around saying that Netscape 3 and Java 1.0 makes the OS iiirrrrrelevant, when if reality what they had was a POS and we still are years away from a distributed, secure environment. It's only natural for a nailhead like Ballmer to look at the 1998 state of Java and dismiss it because it had little business use for Microsoft. --
This could potentially be a step further than "Can we get them to rent instead of steal?"...
Microsoft has the big OEM market by the balls to the extent that they could force them to ship Windows with a bunch of "Windows Installer" hooks for Office tied into the OS. (For those who haven't seen it, the Windows Installer has a way of registring a program icon *before* the program has been installed. Lanuching the icon brings up the installer/credit card entry form.)
Then, not only does every version of Windows carry a heavy incentive for users to buy into MS Office, it could actually be used to prevent unauthorized MS Office installations. Pop-in the Office CD, and have your CD-key checked automagically checked against the MS database. Defeatable? Probably for a knowedgable person. Scary enough to get users to rent the applications? More likely.
(The backlash over disk copy protection in the 80s scared Lotus and WordPerfect straight. It's odd that Microsoft would think that Internet copy protection could actually be reliable enough to not piss their customers off.
Furthermore, if those greedy bastards aren't happy with their P/E ratio now, I don't know what will satisfy them. Help me Government!) --
As a historical note, there were demo clear case Apple ][s and Macintoshes, and if you really look, there were probably 'clear case' IBM mainframes, washing machines, and automobiles going way back. Don't believe the hype. --
Actually, Internet Explorer has the ability to enable JavaScript (and ActiveX, etc) only for Intranet sites, and disable the risky stuff for all others. This can be all under central administrative control. (Mozilla 5 will apparently do the same thing.) Just because Netscape 4.x is feature defecient in this regard, it doesn't mean that Javascript can't be enabled safely in other situations.
The original poster is correct that the real sexy DHTML stuff is only really practical on an Intranet where you have complete control over the client browsers.
In the 20+ years I've been in the industry most serious computer companies have had services as a major line item on their balance sheets.
Just as a point of interest, Microsoft Consulting is a break-even operation by design, and MS makes horks of money with no real services operation. (Novell operates in a similar manner.) This is one big reason there's lots of "MSCE Drones" out there - Microsoft shares the services pie, and generates alot of loyalty out of people and ISV's making their living on MS stuff.
On the other hand, if you've ever been in a competitive services situation with you're traditional minicomputer operations like IBM or DEC Consulting (or others), you know that they'll win most everytime because they can throw licence and hardware discounts at the customer, and tend to cripple the products to make to make value-add services seem more enticing.
I don't know about SCO in particular, but there does seem to be certain loyal ISV base out there (looking closely at Linux). Perhaps they're going to screw their 'friends' and try to monopolize the SCO consulting market and maybe pick up some Linux business too.
(BTW, we'll really see how chummy-chummy RedHat, VA, and IBM are when they start underbidding each other for big Linux service deals.) --
While I'm sure Apple policy and Jobs-Gasse personality issues have something to do with it, in this case it probably makes more sense to follow the money (flowing from Intel to Be).
My suggestion for those who want to see BeOS/PPC live on is to call IBM and suggest that they make a nice fat investment in Be, Inc. They've got the money and it prevents a further weaking of the PowerPC platform. --
As someone else mentioned, an open PPC design is really nothing new - IBM's had one available for years. Think of this move as a little marketing to sell it to the Linux community.
The original intention of CHRP was to provide a mass market computing platform to compete with Intel. IBM would have made their money off the PPC and the chipsets, not the board design. Unfortunately, there's been very little demand for NT, AIX, vapor OS/2, vapor Solaris, and vapor NetWare running on commodity PPC systems. The Linux 'hobbyist' market could revive this. (Anyone have an estimate on Linux on Alpha system sales?)
Slashdot Mirror
Actually, Windows 95 (which was delayed over a year) was suprisingly non-buggy, considering it was built on top of the horrific Windows 3.1. It was no NT/OS2/Unix, but it wasn't intended to be either.
Of course, 4 years of accumulated crap has pretty much got Windows 9x back to the stability level of Win3.1. Time to start over again!
--
It really depends what you run (or don't run). It's not uncommon for (non-MS) Lotus Domino to be up forever on NT, while (MS) Exchange certainly has it's difficulties.
--
Despite it's problems, NT4 is the "NetWare 3" of the late 90s LAN - the entrenched market leader. (Before anyone makes any comparison, NetWare 3 wins in many respects, but I remember quite a few abends and configuration nightmares.)
Windows 2000 probably will be a compelling workstation for shops struggling with crappy Win9x, but the server version includes a number of complex and difficult to understand network services aimed at the enterprise market.
This poses a big problem: Larger shops won't want to move to Active Directory (etc) for a while until they've tested the hell out of it. Upgrading is no longer a tactical decision (like NT3.51 to NT4 or NT4 to Linux/Samba), and more of a strategic one which will involve considerable planning and budget and the typical interminable big-IS project BS.
(I don't see much good information from MS about how an Domain to AD transition is really supposed to work. They probably don't know themselves. Furthermore, they've magnified the problem by unifying the Exchange and NOS directories, which makes the prominence of fucking up an order of magnitude greater. Maybe I'm missing something, but I also don't see a way to run Win2000 server in 'Domain-emulation' mode.)
Smaller shops (which by-in-large are running NT with loosely pieced together WINS systems and broadcasts) don't want or need Active Directory, et al, and will have considerable fear and uncertainty over an upgrade. Everything is kinda-sorta working -- why break it.
All of this adds up to a lot of fear and confusion for NT shops. Add in an economic downturn, and IT budget cuts will probably stagante any new project or make cheaper alternatives (ahem) look more appealing.
So, NT5/Win2000 has a big chance at being the "NetWare 4" of the early 00s LAN -- A good portion of the installed base might just skip the upgrade or switch to a less complex and cheaper alternative.
--
I hope either Queen ("One Vision") or the Nazi party got royalties.
--
It's true that people are happy with the workstation aspects of Win2000, but apparently the server stuff (Intellimirror, Active Directory) is still having some issues in a larger environment.
One possiblity is that Win2000 Workstation will be out soon, but the server version will be delayed by a couple months, and the "Datacenter" server (which supposedly supports 32-way SMP), much later.
--
Just to clarify what I said - binary font files are distributed under a software licence (GPL, MS EULA, X, etc.), just like any other software. The actual art design of the font can't be copyrighted however.
--
Just to make it clear, binary font files still are copyrighted, only the 'design' can not. You can't legally cat YourFont.ttf > MyNewFont.ttf (of course you would need to change some metadata too.)
You can 'copy' someone elses fonts as long as your font drawings are original. For example, you can scan in text and then draw your own outlines around the pictures of the glyphs.
--
Well someone mentioned above that Apple has been folding PPC changes into gcc.
--
As has been pointed out many times, the *most* important applications to many businesses are the ones that have been developed in house in VB/Delphi/VC++/Access/DBase/Whatever. The likelyhood that corporations would/could port these apps to Linux is pretty low.
In house apps nail the average corporate desktop to Windows, so without something like WINE, you'd probably never see Linux on an average corporate desktop.
(The good thing about most corporate apps is that they're unlikey to use the latest Windows voodoo API, so there's no worry about MS breaking WINE compatiblity.)
--
Well, it might mean that gcc has better optimization on x86/Merced than it does on PPC.
--
There's something to your rant. The other day I popped into comp.os.windows-nt.advocacy (for the first time since the good ol' OS/2 days in '94 + '95), and the on-going Linux/NT was a hellava more intelligent and cordial than a typical Windows NT thread on slashdot. Amazingly, less ads too.
--
You said it. Lord knows how many NT admins aren't even aware that it's possible to use a DOS boot disk with their machines. (And with Linux Samba servers, btw.)
--
Well, Linux might not be competitive with Sun/Solaris server systems yet, but it Linux on commodity x86 systems certainly competes favorably with Sun's (and HP, IBM, etc.) traditional Unix workstation business. SGI has figured this out, which is why they're moving away from Irix/MIPS workstations and towards Linux/x86.
In fact, I would guess that most real Linux desktop deployments (aka real work, and not futzing with HTTP/MP3/IRC) are at the expense of a commercial Unix vendor, and not at the expense of Windows NT.
--
ZDNN: So when will you consider that you've 'won'?
Well, when Linux's server market share goes over fifty per cent -- or when Microsoft's stock price crashes, whichever comes first. And again let me interpret that: I'm not saying that Microsoft's stock price crashing is the goal, but when it does, the resistance to doing things right goes away.
We're not taking seats away from Unix, and we're not particularly interested in taking seats away from Unix. Most of the surveys show we're taking seats away from NT -- and I believe Novell has taken a big hit. But we're not killing Solaris. That's OK -- we don't want to kill Solaris.
Since when does Microsoft's stock price have anything to do with the technical goals of Linux? Does RedHat's fat stock increase "resistance" to commercial Unix?
Also, "We're not taking seats away from Unix" !?
The hell you're not. How ESR belives that a free unix is not going to take seats away from $$Unix is beyond me.
"We don't want to kill Solaris" - I thought that creating an Open Source clone of Solaris was pretty much the primary design goal of Linux.
--
Well, to be fair, you had Andresson and others running around saying that Netscape 3 and Java 1.0 makes the OS iiirrrrrelevant, when if reality what they had was a POS and we still are years away from a distributed, secure environment. It's only natural for a nailhead like Ballmer to look at the 1998 state of Java and dismiss it because it had little business use for Microsoft.
--
This could potentially be a step further than "Can we get them to rent instead of steal?"
Microsoft has the big OEM market by the balls to the extent that they could force them to ship Windows with a bunch of "Windows Installer" hooks for Office tied into the OS. (For those who haven't seen it, the Windows Installer has a way of registring a program icon *before* the program has been installed. Lanuching the icon brings up the installer/credit card entry form.)
Then, not only does every version of Windows carry a heavy incentive for users to buy into MS Office, it could actually be used to prevent unauthorized MS Office installations. Pop-in the Office CD, and have your CD-key checked automagically checked against the MS database. Defeatable? Probably for a knowedgable person. Scary enough to get users to rent the applications? More likely.
(The backlash over disk copy protection in the 80s scared Lotus and WordPerfect straight. It's odd that Microsoft would think that Internet copy protection could actually be reliable enough to not piss their customers off.
Furthermore, if those greedy bastards aren't happy with their P/E ratio now, I don't know what will satisfy them. Help me Government!)
--
So one should worry that companies are using their Intranet web servers to steal "assets" from their own computers? OK.
--
As a historical note, there were demo clear case Apple ][s and Macintoshes, and if you really look, there were probably 'clear case' IBM mainframes, washing machines, and automobiles going way back. Don't believe the hype.
--
If one wants to pay a lot of money for a black system, they can always call IBM.
--
Actually, Internet Explorer has the ability to enable JavaScript (and ActiveX, etc) only for Intranet sites, and disable the risky stuff for all others. This can be all under central administrative control. (Mozilla 5 will apparently do the same thing.) Just because Netscape 4.x is feature defecient in this regard, it doesn't mean that Javascript can't be enabled safely in other situations.
The original poster is correct that the real sexy DHTML stuff is only really practical on an Intranet where you have complete control over the client browsers.
--
Honest question - I thought most modern mainframes could run TCP/IP and the older SNA protocols were encapsulated.
--
In the 20+ years I've been in the industry most serious computer companies have had services as a major line item on their balance sheets.
Just as a point of interest, Microsoft Consulting is a break-even operation by design, and MS makes horks of money with no real services operation. (Novell operates in a similar manner.) This is one big reason there's lots of "MSCE Drones" out there - Microsoft shares the services pie, and generates alot of loyalty out of people and ISV's making their living on MS stuff.
On the other hand, if you've ever been in a competitive services situation with you're traditional minicomputer operations like IBM or DEC Consulting (or others), you know that they'll win most everytime because they can throw licence and hardware discounts at the customer, and tend to cripple the products to make to make value-add services seem more enticing.
I don't know about SCO in particular, but there does seem to be certain loyal ISV base out there (looking closely at Linux). Perhaps they're going to screw their 'friends' and try to monopolize the SCO consulting market and maybe pick up some Linux business too.
(BTW, we'll really see how chummy-chummy RedHat, VA, and IBM are when they start underbidding each other for big Linux service deals.)
--
While I'm sure Apple policy and Jobs-Gasse personality issues have something to do with it, in this case it probably makes more sense to follow the money (flowing from Intel to Be).
My suggestion for those who want to see BeOS/PPC live on is to call IBM and suggest that they make a nice fat investment in Be, Inc. They've got the money and it prevents a further weaking of the PowerPC platform.
--
I don't know about your setup, but NS 4.61/WinNT leaks memory like a sieve on my box.
--
As someone else mentioned, an open PPC design is really nothing new - IBM's had one available for years. Think of this move as a little marketing to sell it to the Linux community.
The original intention of CHRP was to provide a mass market computing platform to compete with Intel. IBM would have made their money off the PPC and the chipsets, not the board design. Unfortunately, there's been very little demand for NT, AIX, vapor OS/2, vapor Solaris, and vapor NetWare running on commodity PPC systems. The Linux 'hobbyist' market could revive this. (Anyone have an estimate on Linux on Alpha system sales?)
--