Editor's Note: The Readers Respond
Going Through the Mailbag
Last week's editorial on Microsoft and Linux provoked the most responses via e-mail (as well as responses posted on Linux Today) in the history of LinuxPlanet. Since we don't have talkbacks on LinuxPlanet yet (they are in the works, however), many of you reading the column via e-mail had no public mechanism for responding to the column--so you did the next best thing and deluged me with e-mail. Virtually all of the responses were intelligent and made for interesting reading, even (and probably especially) those that disagreed with what I wrote. Featured here are the best of the responses.
There is one thing, however, that all the letter writers wrongly assumed: that when I spoke of the browser of the future, I was speaking of Mozilla. Not necessarily. For the record, my personal viewpoint is that within the next two or three years (the timeframe used in the original column), the Konqueror Web browser will be a major force in the Linux world. We are seeing the power of the Open Source movement, and it won't be long before projects developed under the KDE and GNOME umbrellas will surpass commercial software.
>Let's face it: if you were doing some mission-critical work
>via a Web browser and needed an OS that provided a high
>level of reliability, which would you choose: Linux or
>Windows? There's just no comparison.
You're right. As much as I love Linux, I have to admit that Internet Explorer 5.5 is lightyears ahead of Netscape, Mozilla, Opera, and Konquerer in terms of speed, stability and features. I would never run a web SERVER on Windows, but if I were doing some mission-critical work via a web browser, I'd be forced to choose IE5.5.
As far as 'high levels of reliability,' Netscape for Linux just isn't there. Mozilla is just as far away.
In your column, EDITOR'S NOTE, in the June 29, 2000, issue of LinuxPlanet Newsletter, you said that the reason for the Microsoft.Net initiative is that Microsoft is following the market to Web-enabled applications and that that signals the death knell for Windows.
That may well be true, and I certainly don't have any inside information, but I have an alternative theory.
Microsoft currently owns the desktop market, but it seems to be losing to Linux in the web server market. Linux can do well in the server market because it is more economical and more reliable than Windows, and because web servers serve open-standard HTML pages and can therefore run any OS. Microsoft.Net applications will certainly require a server running Windows. If Microsoft can get enough key web sites to serve Microsoft applications, competition will force other sites to have this capability, too. Even if sites never use the capability, FUD will encourage them to run Windows on their servers, just in case.
Sun purchased StarOffice supposedly to sell more servers. I think Microsoft is developing Windows.Net to capture the server OS market.
With Microsoft owning both the browser and server markets, it would have no need to support open standards. It could dictate the formats for data exchange on the web. The only platform capable of fully browsing the web would be Windows. Netscape Navigator, Opera, lynx, etc., would become barely usable, and consequently, so would Linux as a desktop OS.
That's my theory, anyway.
I just read your article. I go along with your main points, but I'm baffled by the idea of moving applications off the desktop in the first place. To me, it seems like a exercise in radically decreasing system reliability for trivial benefits. In short, I don't see why anyone would willingly consent to adopt that model.
A single stand-alone workstation is pretty reliable, at least if you're running any kind of Unix-oid op system on it. And if it does fail, only one user is affected. That limits the lost time when the inevitable finally happens.
On the other hand, consider the application server that maybe a hundred users are logged in on. Let that crash, or get cracked, or slow down from overload, and a whole department or a whole company is brought to a screeching halt. Now consider every router, firewall, hub, name server, and piece of cable between those users and the server. Every one of those is a single point of failure that can shut everybody down. And if the application server isn't in-house, but belongs to an ISP, or even some service provider on the opposite side of the world, well, we all know how clogged the Internet gets sometimes.
Yes, sysadmins like loading and configuring and application once and letting everybody use it. However, pushing it to multiple destinations over ftp isn't a heck of a lot different. And then, lots of users need to customize it anyway, so all their config files are going to have to be maintained and backed up.
Today's office software is a per-seat expense, and that's a motivation for minimizing copies. At the moment. But competition has set in among the commercial office suites. And then there's Gnumeric, GIMP, and the up-and-coming AbiWord. License cost isn't going to be a decision criterion for long.
30 years ago there would have been an economic argument for saving hardware cost on each user's desktop -- a central server requires only one set of hardware capable of storing and running the application software. Today, of course, a decent monitor is 3/4 of the cost of a supercomputer, so nobody's even interested in booting diskless any more, and I don't think anyone even sells a desktop machine with less than 64 MB of RAM.
Personally, I think Microsoft's latest move is just another instance of rent-seeking. This time I don't think they'll be able to make it stick. The world has passed them by.
I enjoyed your comments in today's issue. Even when I don't agree, you make me think.
A couple of comments. You mention making the move from Windows or Mac to Linux easier. Keep in mind what is coming on the Mac front. The kernel of MacOS X is BSD. Further, I've heard that there will be (already are) one or more Linux distributions running on the Mac G4 so I presume you are talking about moving from MacOS 9 to Linux.
I also noted your comments about web-served applications. I have tried out a few for our business and am mightily under whelmed. I know that they will improve, that cheap and plentiful bandwidth is just around the corner and that network latency is much over-rated but I still have some reservations about this ASP juggernaut.
Robert W. Meyer, D.Sc.
regarding your editoral, specifically:
It also signals the death knell for Windows. Let's face it: if you were doing some mission-critical work via a Web browser and needed an OS that provided a high level of reliability, which would you choose: Linux or Windows? There's just no comparison.
considering IE's superiority, it's lack of availability for linux, and the application being hosted on a server (therefore, desktop stability is irrelevant), i would choose windows (or mac, or any platform where IE was available).
do not underestimate the subtlety of microsoft's business strategy.
p.z. linuxplanet should provide a mechanism for publically posting comments on articles, e.g., your editoral note.
Greetings and Salutations,
In your article "Moving Applications to the Web" published on linuxplanet you concluded with:
"Let's face it: if you were doing some mission-critical work via a Web browser and needed an OS that provided a high level of reliability, which would you choose: Linux or Windows?"
Alas, the biggest problem with that is finding a robust web-browser. I expect IE and Netscape to be flaky and unreliable, but unfortunately, Netscape under Linux is just as bad (if not worse) than the doze port.
My experience with Netscape under RedHat and Debian is poor at best. If I use it on mostly text website, it might go several days before going belly up... if I use HEAVY graphic sites, maybe one crash a day... if I enter Java heavy sites, it can crash several times in a day.
After a number of discussions at our local Linux User Group (NWCLUG - NorthWest Chicagoland Linux User Group), I've found my experiences to be the norm rather than the exception.
Also keep in mind, Microsoft loves its active-x files... no matter how you slice it, they wont run on my Linux box...
All things considered, I don't see the movement of MS to web-interface making one bit of difference to non-ms platforms.