Fog Creek Software
Discussion Board

Is the tide now going against Open source?

It is interesting, but about 2, or 3 years ago, it seemed that open source was going to change the world as we know it.

Companies like Oracle, Sun, and the incredible IBM were all making waves, and jumping on the Open source band wagon. Announcements were occurring daily.

However, I see a few trends, and a few problems right now. In fact, I am not only seeing Open Source loosing steam, but I am now seeing a backlash in the industry against OS.

I am only being a messenger  here, but I am getting a lot of feedback, and much of it is not very good for the OS community After all, business has the FINAL SAY on how successful software will be.

I also should state that I do have software of mine running on Linux. (in USA, France, Italy, Germany, Canada. and likely a few other Countries I am not remembering just right now)

I am not a dreamer, nor do I work on “hope” that things should be the way we WANT  them to be. I try to see things as they are.

Anyway, to me, what is REALLY important is the kind of feedback I am getting from companies and business that are using Open Software (OS) systems like Linux.

There is some real serious problems.

The first one is I am seeing is usability, and ease of use. Right now, some of the open source stuff is really starting to show its weak spots. You see, now that some companies were sold the open source bills of goods, then they now are also are now seeing the downsides.

I was at the Tech Ed show in Edmonton on Wednesday, and during the lunch break I was talking to a long time client of mine. He is an accounting guy that over the years hired developers like me to help him build and integrate his accounting systems into vertical markets that he sells into. He is a typical VAR type software vendor. This VAR/owner of this company also does a good portion of software development himself. However, he is more accountant then hard core developer. Hence, this VAR also brings in subcontractors guys like me from time to time to help him with the more difficult coding stuff..

Anyway, he was complaining to me as to how hard some system administration stuff was on his Linux box. In fact, he actually said he wished he had went with windows NT (aka windows 2000 server). In his case, the D3 raining database system that his accounting software runs on can run on windows NT, or run on Linux. He went the Linux way.

Some of my software also runs on Raining Data’s D3 multi-valued database system, and thus explains why my software is running on some Linux boxes around the world.

Anyway, the first problem or complaint that this VAR guy had to me was his Linux support guy is ALWAYS coming in to tweak a whole bunch of settings (and, lack of standard settings for how things should be done seems to come up here!). Worse, is even simple things like setting up printers on the network requires this var to bring in a support guy.

In other words, what is nearly automatic, and trouble free in windows is a royal pain for this var guy to manage. This is not really a question of having a good Linux admin, but the fact of setting up printers and stuff like that SHOULD NOT cause this fairly computer literate person to get stuck (and have to bring in a Linux pro all the time). He now wishes he went with NT.

Of course, making this situation worse was here we are are Tech Ed watching some of the tech net demos and how MS server administration is SO easy as compared to Linux.

So much so, that this guy is cringing. Without question, having known this guy for years, he would have little trouble with setting up printers for a windows server. (I sure I could too). However, setting up printers and drivers for Linux? (you are kidding ..right?).

This is a typical business environment. There are a lot of companies in the 1 to 25 pc range that do NOT have a full time admin. These companies using Linux are experiencing more difficulty then they should. Further, the cost of finding people to come in and setup a printer etc is MORE then what the windows side is (lots more people can set a printer in windows!).

Further, with stuff like the SBS (small business server…a instant business in a box server from MS, this case is even FURTHER in favor of using MS. You see, for these types of business, you got a typical server, and anywhere from 5 to 25 windows boxes. A good number of business of course go with a Linux box, but run windows clients (you got that killer application called samba, and real nice box of goodies do come with Linux). Up to this point, things seem ok with either choice.

However, now that MS is pre-packaging the small business server, then you get virtually everything you need to run the business in a box. In fact, you get a sweet break on MS pricing of exchange server, and a bunch of other stuff.

What is incredible is that this kind of offering was not done sooner! I mean, it has been, what 6 or more years that we purchase pc’s with the OS installed? In fact, when I look at the cost of SBS, you actually save A huge whack of money since you don’t even have to bring in someone to install, and setup a TON of software (the price break + not having to install and setup a huge amount software makes so much sense. Just saving the installing, and initial setup of all this software is worth a lot in terms of INITIAL support costs. MS really should call this Instant business in a box server.

Of course, a business does NOT setup a server everyday, so this cost issue might not be too great. However, anyone who can turn on a pc can get SBS running. However, what is even MORE important here is the new CLIENT PC administration that these systems offer. You see, today, much of the cost of running a small business system is keeping the CLIENT PC’s updated.

With SBS you can now manage software updates, and even software installs.

I mean, why in the year 2004 do we actually send a tech around to each pc to install a new version of office, or stuff like a new version of  Simply Accounting clients on each pc?

In other words, MS is doing what it does best, and is leveraging their server products to work WITH THE windows desktop. This makes management of client pc’s VERY easy. This stuff is getting VERY slick.  So, you got 35 pc’s hooked up to the server.  You need to roll out new Excel, or whatever to the sales department.

However, does your server let YOU manage all the pc’s, group polices, AND software installs? And virtually everything else you need to do from a central location?

Hum, you still want to run Linux and samba now? Forget it! That Linux box will do NOTHING to reduce your client pc support costs. MS is going to kill Linux in the small business service market with this product. It is so slick, and it is easy to use.  It is all pre-installed (sql server, web services, software updates, exchange server (mail etc. etc.)…the whole thing is “ready” to go. That var guy of mine would have been “ready” to go if he went with a sbs box.

The real key here is the client pc management stuff. I will say that the pc management stuff is not quite as good as the big corporate edition of sms from Microsoft (sms = system management sever = a system to install and maintain software to all client pc’s in a large company. This system also manages windows update, and can save internet bandwidth by having client pc’s get their updates from internal servers..and not ms).

So, the SBS does not have full  SMS, but it sure the hell is nice. You get the ability to update software on each pc in your office. (good bye to that tech guy that runs around from pc to pc to pc to pc to pc to pc to pc to pc…).

So, a real large portion of your costs are NOT the server, but CLIENT pc support. Products like SBS are winning big time. You have no idea how slick, and how cool some of the system management goodies that has been setup on SBS. Further, even general system admin stuff has been cleaned up, and made VERY easy to use with a lot more wizards. This means that those computer savvy business owners (or self appointed tech support guy from the accounting department can do this stuff).

For sure, for more difficult setup stuff, these companies will STILL call in their tech support guy, but MOST of the time you can do this stuff on your own. (and, there is a heck of a lot more support guys in the yellow pages for windows if you need them anyway).

And, I know for a fact that this VAR friend of mine I was talking to could EASLY install and setup and maintain printers for his network if he had SBS.  Right now, he is not happy, and is spending WAY too much on support (he now regrets having chosen Linux). For those companies that got sold Linux to save money, they are now seeing REAL increases in software support costs.

So, I am touching on two VERY important issues here:

1)    does your SERVER system have a GREAT bunch of pc client management tools built in?

2)    Can someone with not super server admin skills setup a printer and other stuff?

And, before anyone gets on my case, you all know who Eric S. Raymond is. right? I have his book..and in fact purchased 3 hard copies. They are all currently lent out right now to my friends. To anyone who I lent the books….if  you are reading this…I want my copes back!!!...they are hard cover editions!

You can read what Eric has to say about setting up printers. and how he BLASTS the Linux community here:

However, I had the above feed back long BEFORE I read Eric’s article. In other words, REAL business people I am talking to are now complaining about how difficult it is to do basic system admin stuff on a Linux box. They are complaining how they have to ALWAYS bring in tech support for stuff that is dead simple with a windows server.

I remember how hard networking used to be back in the pre-windows days. However, this hardness sure made Novell a lot of money. I can now setup a small network of pc’s in my sleep. I assure you I am only some dumb developer, and not even close to being a network support guy…but that stuff is really automatic right now in windows.

And, that Linux box??? That Linux box does NOTHING to save on client pc support (this is where SBS is really winning the battle).

The next issue is one of help systems, and also one of user feed back:

MS is now integrating their help systems with the on line communities. On-line help, and even systems to work with newsgroups are actually being integrated into MS help systems. Office is first, and other products are to follow (default help in office 2003 is on-line). What this means is that MS is using USER feedback (not developers ) to figure out where usability problems exist in their products. I can say from personal experience, the results in office 2003 are simply astounding. (the help for some stuff I used has NEVER been better, where as previous editions sucked).

Again, this is all about the user experience, and how you can make things easy for users. It is no surprise that MS is MUCH better then Open Source for usability. However, MS is accelerating the use of on-line communities and integrated help systems into their products. What is critical about this process is that feedback is being feed to their usability experts on how to make their products work better. This usability system is much like the feedback system that open source uses to fix bugs. MS is thus harnessing the user communities out there in a MUCH better fashion then the open source guys. In fact, MS is using the VERY SAME model that makes Open Source possible! What is different here, is that MS is applying these concepts of connectivity to end users…and not ONLY to developers.

Who the heck is going to start gathering these usability issues, and customer support issues for Linux? (they better wake up. as this is so important, that I see battle after battle being lost unless this issue is addressed by the open source community).

And, MS is doing this help stuff with a well thought out plan. In my visits to Redmond, and seeing the sophistication on how they are integrating the internet, and customer support INTO their products, I am now actually beginning to realize that open source is not even in the same league in terms of delivering consumer products. MS is so out maneuvering the open source people by a wide margin on their own game of using the connectives of the communities at large. Open source seems to have zero response to this issues, and then wonder why they can’t make inroads into consumer software?

The next issue I am going to deal with is the issue of software quality. To me, the Open Source model of bug fixing is rather IMPRESSIVE. Release often, and fix bugs in a rapid fashion. Further, the speed of software development is not hurried, and usually results in very high quality products. Bugs can rapidly be fixed by developers around the world (how the updates will get back down to the consumer is still a big issue. And, while developers can work on bug fixes…who, and how does a consumer of this software tell about the bugs? MS is again that MS is doing a better job on this feed back stuff that OS invented!).

What about MS in this regards, and in terms of software quality? I remember reading stories about a us citizen called W. Edwards Deming. Mr. Deming is often credited with the rise of the auto industry in Japan. This man brought statistical quality control to the Japan auto industry (the north American industry did not listen to him….well, they now use all of his ideas in quality control to day).

MS right now is going through a revolution that is likely LARGER and more important to the software industry then what the auto industry experienced.

I can tell you from first hand when running office 2000 on a winxp box, it was not the most stable product I have used. In fact, I would rate the quality and stability of office 2000 as quite poor.

Then a new release of office xp came out. The improvement of office xp is STUNNING.

Again, I see clients all day, and talk to as many developers and people I can. Virtually all I talk to find a HUGE improvement in the quality of office xp (2002). What the hell happened to make this release SO MUCH better then office 2000?


You see, MS is now integrating Dr Watson error reporting into most products. The results are amazing. Office XP is the first product released from MS that had fixes, and quality improvements based on the “send error reports” that Dr. Watson sends. So, while we might not like Dr Watson…it is a tool that is CONSIDERABLTY being expanded to other products. Soon it will be a standard part of ALL software products.

A graph of errors in MS’s bank of servers that munches and crunches these Watson errors all day will look like:

  | *
  |      *
  |        *
  |        *
  |          *
  |            *
  |                *
  |                          *
  |                                *
  |                                        *

The above graph is the kind of bug errors graphs the support teams look at all day. The most common and largest bugs are on the left. So, what happens if we fix the first 8 most common bugs in the stats curve above?  Well, you don’t have much of a cure left to plot do you? Well, actually what happens is that they chop out the first 8 bugs fixed, and then you stretch out the curve to take up the above grab space, and then re-draw the curve again. You now get the same curve again, but the bugs you are fixing are of course of a much lower frequency. You attack the bugs at the top of curve again. The next graph you draw will look VERY MUCH the same as above. You do this over and over (this is actually what MS does in this case!)

So, the curve looks the same. but you are working on a smaller set of bugs.  You keep hammering this curve down, and all those pc’s all over the world are sending your HUGE bank of servers bug info in real time.

In fact, MS will now even inform some software vendors that THEIR software is causing stability problems on windows! (again, who the heck is going to do that for open source?).

I can tell you right now, that the quality of new products is soaring at MS. The increase in stability, and reduced number of problems is simply amazing here. The improvements I see in terms of this process is very noticeable to me in office xp.

I am also seeing this in office 2003 now. This once again shows how MS is beating the open source people at their own game (that game is one of using users to gather bugs and problems, and then feeding them to a central location). It is just that MS has automated these software bugs systems right into their products, and again open source is being left so far  behind.

Where is the automated bug reporting systems for open source software? And, a even larger problem exists as to where will the bugs be sent? And who is the team that spends on all day munching and crunching these numbers with the huge bank of servers required to support the volumes that dr Watson generates?

These built in feedback and error reporting systems is creating a revolution at ms, as they can now instantly see what the hell happened (why are people getting a gpf?). When MS has a problem, then they know real fast! Even more incredible, this bug process functions faster then what humans can do to gather these problems. Further, the REAL big issue here is how software interacts with other software. You might have a system problem, but that problem is ONLY due to the fact that you go a old version of sql server, and are running some weird virus protection system. MS plans even further bug integration into their systems.

I see this as a quality revolution at least as large as what Mr. Deming did for the Japan auto industry. We are JUST now experiencing the first batch of products based on this process. In a few years, competitors that don’t do this will be left even further behind. This whole process can’t be built over night….and has taken MS quite a while to build.

(this is a amazing lesson in software quality control).

Another issue here is use of standards:

One of real sticking points with propriety products like Office/Word has been the use of propriety document standards. While the open source community is STILL arguing what format to use, all of office 2003 has great XML support. You can now create a cool looking formatted word document on a 1981 ibm pc using gw-basic, and simply write out XML…word will load it. (and wow. does it work great?). Have you guys seen the cool xml support in word and Excel. Once again, MS is taking a open text format. and running with it. Many have commented that Open Office looks a bit old, and tired already (in the range of the older office 97).

The real problem I see here is that the gap in terms of help, support, and new features like XML support is actually means that ms Office is stretching out the lead over Open office. By the way, what great new features does Open office have? (and, how does the help and customer feed back system work? And, what is open source doing with open office in terms of the REAL NEED to secure company documents? I just see the gap getting wider between MS Office, and open office. That was NOT the case 3 years ago.

Open office seems to have lost it direction, and does not seem to know what road, or where the product should go. How can Open Office know where to go without a strong company committed to its future direction? Who is going add digital rights management to open office (ms has the plugs in now. and MANY companies are just waiting to have protected documents that you can’t take home, or email to competitors down the street (you might not like digital rights management, but I know a lot of companies that will go ape over the ability to protect their documents. Companies used to be able to protect their documents by limiting who had keys to the file cabinets). So, now we have Point Share services from MS that does this (and, it also integrates a whole bunch of workflow management into office also).

Worse, is now the deal between MS and Sun means that Sun wants to distance it self from open office. I think most would agree that Sun opened sourced office as a tactic to just hurt MS and the office/windows franchise. Sun has NO intention of battling MS anymore. And, they have given up on this lets try and make MS hurt and bleed concept.

Sun has too many troubles of its own, and it now needs, and wants peace with MS.

The Open source community does loose with the MS/sun deal as sun is not so interested in Open Office as it once was. Hum, it looks like IBM might step in and take over Open Office (gee, what features did open office get via open souce after sun purchased the product anyway?).

Mean while, while all this stumbling in open office is occurring, MS office just zooms aways with new great xml stuff, and the new digial rights via share point. And, also, lets not forget all the integrated help and bug feedback systems that MS now has in office to make the product even better next time around.

So, just as people are starting to look at Open Office, things for the product don’t look that good, and Sun is not going to be much help anymore on open office.

Between systems that support the desktop, and how MS is levering the MANAGEMENT of pc by their servers, and how they are using the Internet to improve both their quality, and the “users” experience is a long term issue..and will have long term effects in how the next versions of these software systems look.

In the late 1970’s, things like structured programming came along to deal with the complexity of software. Then stuff like OO came along. Each time it allowed us to build more and more complex systems. Today, it is automated user feedback systems for BOTH bugs, and MORE importantly, feedback systems for USERS that use the product. These two concepts that MS is adopting is very much like Mr. Denning approach to quality control.

And, more important, these tools allow ms to make the NEXT generation of software MORE complex then we can imagine now. And, look, the ms .net stuff is now finally coming of age.

The open source community now needs a new model, or some serous updating to their existing model in terms of user feedback, and automated bug systems. The current open source model is being pummeled by MS right now.

Times are changing and what worked 5 years ago for development of software will NOT work now. The industry is adopting new tools now just like we always did (structured program, oo etc).

The general software community must start to adopt these new systems of integrating the customer into the software process, or they will loose big time.

MS is now hijacking the BEST ideas that open source offers in terms of user communities…and running with it….

Once again competitors to MS are making really big mistakes….

Albert D. Kallal
Edmonton, Alberta Canada

Albert D. Kallal
Sunday, June 6, 2004

Sure, opensource is largely bullshit, and mainly exists because people aren't ready for Gnu.

OTOH, putting software companies out of business is a Microsoft yardstick of success. Many "opensource" groups have already claimed victory by accomplishing their goals. IBM claims to have profited from its opensource investments. Gnu claims wild success, and counts backlash as proof. Many skilled users never spend a penny on basic software like OSes and programming environments.

Gnu and most opensource are making Microsoft bleed by accident. And it makes sense; Microsoft killed so many companies that the super-viruses are left, which can't really be killed.

Tayssir John Gabbour
Sunday, June 6, 2004

WOW!  That has got to be your longest ever post Albert.  I enjoyed reading it.  I especially liked "While the open source community is STILL arguing what format to use"  Well, their Unix guys.  Arguing is second nature.

Myself I am just wrapping up Mark Minasi's "The Software Conspiracy:  Why Software Firms Create Faulty Products, How They Can Harm You, And What You Can Do About It"  It is a very interesting read.  Basically the pc revolution is winding down.  The leap from Dos to Windows was greater than from 95 to 98, etc.  We are getting to the point where the best features to a lot of types of software are already discovered.  According to him we are closing in on the time when competing on quality is going to matter.  Microsoft does seem to be building in the architecture to do that. 

I also agree that most small businesses without a full time admin would be better off with Windows products, assuming they have someone qualified set up things like the servers and virus protection.  You shouldn't need an admin to add a printer.  On linux you need someone who can write printer drivers it seems.

Sunday, June 6, 2004

Well, I'm still of the opinion that it all depends on what you're used to.  You've brought up alot of anecdotal evidence - this and that - of how Linux is hard, yadda, but nothing specific.

I'm integrating code into a 3rd party's VxWorks develpment environment, and they insist on doing development in Windows.  Fine.  It took them an entire day, expending the resources of several embedded software developers just to get VxWorks build to work on our code.

Obscure nonsense?  Hell yes.  Specific to my situation? Absolutely.  Much the same way as your Raining Data yadda, yadda, Linux tweaking app setting horror story goes.


"However, does your server let YOU manage all the pc’s, group polices, AND software installs? And virtually everything else you need to do from a central location?"

HELLS BELLS!  I can't even set up a reasonable policy such that I can permit my kids to run games and NOT install the latest virus/spyware across the entire friggin user/admin space of the game PC.

"you still want to run Linux and samba now?"

Do it all the time.  Even our church's secretary can handle that.  Why is this hard?  Set and forget.  I can easily ssh in and config it from work when she has questions.

"I can tell you right now, that the quality of new products is soaring at MS. The increase in stability, and reduced number of problems is simply amazing here. The improvements I see in terms of this process is very noticeable to me in office xp."

Office is 10 years  old for crying out loud.  And its getting really, really good now!!  Woohoo, have a party, let's go to Whistler!

Fact is Microsoft is struggling outside Office/OS.  Your particular view is very narrow, and perhaps in this specific view, Microsoft looks awesome.  I can easily beleive that Microsoft and VARs can deliver great solutions from detist's office to an accounting firm of a couple hundred people.  I can also believe that no one is going to take the time in OpenSource land to create an integrated solution to compete against Microsoft.

It isn't that it cannot be done - it merely is a fact that no one cares.  So what?  Microsoft and its VARs can revel in their glory that this segment is squarely secure and so are their jobs.  Again: goto Whistler and have a party.

"I just see the gap getting wider between MS Office, and open office. That was NOT the case 3 years ago."

Not I.  Our office is fully in the Microsoft camp as far as the office suite goes.  I use OpenOffice (except for one specific instance where tables and numbering were completely hosed, I had to make sure it wasn't OO that was screwed) and no one else knows the difference.  We pass documents via email just like the idiots God created to do (never mind that we could use Perforce, but the Office crowd is TOO DAMN STUPID to figure it out, so we use email).

Albert, your clients are MS centric, and thus so are you.  So it should be.  Software should be driven by requirements, and if MS fits those requirements best, then MS it should be.

Just realize that what helps your clients does fit all.  Additionally, open source software doesn't really care what Microsoft does or doesn't do.  Its a non-issue.  Perhaps Sun microsystems cares, and maybe IBM cares, but open source projects are not aimed at defeating Microsoft - which seems to be a premise of your argument.  Open source provides...  what open source provides.  That's all. Don't like it?  Move alond, nothing to see here.

Sunday, June 6, 2004

Lots of typos.  Hope you can figger the intent by context.  Like does _not_ fit all.

Sunday, June 6, 2004

One thing that open source has going for it is that all the cool kids are doing it.  It seems like the quirky, neat, interesting things that are going on are disproporitonately (fuckit) open source or at least built on open source.  BitTorrent is the first example that comes to mind.

A big thing that MS has against it is the perception, I think increasing, that MS wants you to do and be a certain kind of computer user, and if that's not a good fit for you, too bad.  The unwillingness of MS to put a popup blocker into IE is a handy example.  The distant silhouette of Palladium or whatever it's called nowadays is a more ominous one.

This is not to say that open source doesn't have its problems, and yes ease of use and administration is way at the top.  Most people don't think it's fun to solve a puzzle every time you want to do something new with your computer.

I think the growing commercialization of Linux has some promise there, though I don't really trust Novell to get it right.  On the other hand, it's not like open source platforms are going to go out of business, so maybe the next guy will get it right.  On the third hand, maybe software patents will keep the next guy from even trying.

Matt Conrad
Monday, June 7, 2004

Before I finish reading your post (I'm a slow English reader:), I would like to point out a couple of links (*1) that show that Brazil is starting to support  Open Source on large-scale. President Lula still has 2 years left, before his mandate runs out, after which he will try his second election.
If President Lula wins his second election, it could mean a long Open Source trend in Brazil.
Miguel de Icaza (*2) knows (*3) that there are countries like Brazil where Open Source fits nicely. Miguel de Icaza has travelled to other countries promoting Open Source (*4).

*1 -
*2 -
*3 -
*4 -

PS: Yesterday I didn't know about the latest Open Source initiatives of Brazil. It is a coincidence.

Monday, June 7, 2004

"Where is the automated bug reporting systems for open source software?"

KDE has this.  In fact, I started seeing this dialog popup around the same time as I started seeing it Windows so I don't actually know who came up with it first.

"And, a even larger problem exists as to where will the bugs be sent?"

The KDE team.  They build as many applications for KDE as Microsoft does for Windows.  And they probably pass on bug reports for software they haven't authored.

"And who is the team that spends on all day munching and crunching these numbers with the huge bank of servers required to support the volumes that dr Watson generates?"

They do it -- don't ask me how they do it.  Each version of KDE is more stable and faster than the previous version.

I'm just pointing out one example of where you THINK you see the whole picture but you're only looking at one corner.  In the situations you describe, I wouldn't hesistate to recommend Windows servers.  But then there are other situations (e.g. webserving) where I wouldn't touch Windows with a 10-foot network cable...

Now, I don't have a bias either way.  My desktop machine is Windows XP Professional.  Every other computer in the house is Windows except the firewall.  I do various kinds of development including Visual Basic.  However my main day-job is developing and deploying a large web application in PHP.  I see the value in both environments.  They both have their strengths and weaknesses and perhaps, in the end, neither will dominate the other.

Almost Anonymous
Monday, June 7, 2004

No the tide isn't going against Open source.

There's a fundamental misconception.

Open Source is not an alternative marketing strategy to closed or proprietary source.  Open Source does not mean that you need the source in order to use it, it simply means its available if you do need to use it.

Open Source has many good development teams that use good QA and don't necessarily rely on the many eyes of the Bazaar to fix it.  Automatic reporting procedures are all very well (and some of us have used them for a great many years), but unless you have a significantly large population of users they aren't going to make that much difference over regular QA and beta testing procedures.

Microsoft does have that large population of users, as does any of the significant Open Source products.  Perhaps the only difference is that it's relatively straightforward to know the current outstanding bugs on an Open Source product and a matter of surmise, conjecture, rumour and personal experience as to knowing outstanding bugs on Microsoft's products.

Microsoft is not bad at publishing its bugs but it isn't the almost complete exposure that happens within well run Open Source development groups.

Not that you have to be Open Source in order to have frank disclosure, I maintain bug databases that are open to my users for the products of mine that they have.

Is closed or proprietary software easier to install than Open Source?  Are apples easier to eat than pomegranates?  Is there much point comparing them?

Accounting systems do not particularly lend themselves to Open Source, however they are frequently consultantware (its part of how I earn bread, though not much butter).  Successful consultantware has to be openly configurable and preferably in ways which a properly trained user/administrator can do for themselves.

Perhaps a closer installation comparison would be Open Office against MS Office.  From my last experiences of installing MS Office I'd find it safe to say that Open Office was the simplest to initially install and certainly much the simplest to upgrade.  However, one experience does not scale to millions, nor even the norm and for that reason I'm wary of making claims or even directly comparing them.

But I have two examples of Open Source, that are not Apache or Mozilla but that fulfill niches, one in almost complete isolation and the other in competition with other proprietary and open systems.

ImageMagick is an image file manipulator, to convert transform, merge and generally muck around with images.  It runs on a generality of platforms and is frequently used to provide thumbnails on the fly or publish images with watermarks.

Its available as either source or installable versions.  There is no expectation that anyone needs to change the source but its there if necessary.

You can report bugs and be informed of them via mailing lists and web pages.  There's no integrated automatic call home reporting, but then there's no particular need for it.

The other is analog a web log statistics and reporting tool.  There are a great many log reporting tools, possibly others that are easier to use.  Again its source is available but as all of its behaviour is configurable with switches this is really consultantware.  It would take longer for me to inspect the source and affect its behaviour than it would to experiment with the switches and use the existing documentation.

Again there's a reporting procedure and public bug lists and mailing lists.

There is no tide of usage of Open Source, its just that consultants will pick the best tool for the job, and that's always been true.

Simon Lucy
Monday, June 7, 2004

You raise some good points and obviously know a lot about the MS side of things.

I think a lot of your concerns / points about OS software can be addressed by projects already out there - check out Ximain Red Carpet Enterprise.

I do agree that the OS world does not have many of the turn-key type products available, but I would be very suprised if IBM did not leverage the technology out there and push a bundle sometime in the next couple of years.

Andrew Murray
Monday, June 7, 2004

Wow Albert, that has to set a record as the longest JoS post. In fact, you've written more in that one post than Joel himself has in the last month and a half. No comment, just that observation.

Monday, June 7, 2004


I think your comment shows why open source is a good thing:

"What is incredible is that this kind of offering was not done sooner!"

Would Microsoft have done this if it hadn't been for the OS competition?  I don't think so.

When OS really made the headlines for the managers is when Microsoft started getting heavy handed with the licensing policies.  Would Microsoft have backed down like they did if everybody hadn't started evaluating OpenOffice?

Open Source may not be right for you or a lot of other people.  However, I do think it is good for the market.

Ged Byrne
Monday, June 7, 2004

Ged, I think what Albert was saying is that the argument has moved on from this idealistic clap-trap about markets and that people actually using foss products are starting to query the mythology.

Monday, June 7, 2004

In the end, it all comes down to insentives.
I prefer a company that makes money by selling me boxes of products, and takes service and support as a cost center, over a company that gives me the software and takes it as a loss leader, but makes its money from service and support.
I do not believe in getting anything "for free". Contrary to the latter, the former actually has an inherent incentive to make products better (wether this incentive stays as strong once one moves to a subscription model is another matter).

Just me (Sir to you)
Monday, June 7, 2004

>The unwillingness of MS to put a popup blocker into IE is a handy example.

Hum, actually there is one in the next release. In fact, if you installed the windows XP service pack 2, get this feature...

Albert D. Kallal
Monday, June 7, 2004

I am very glad because I am seeing signs of OSS losing momentum, as in my opinion OSS is BAD for most developers.

OSS creates a business environment where some star developers like Linus make a lot of money, and the rest have to put up with reduced wages.

Michael Schredder
Monday, June 7, 2004

Ho hum.  What's the evidence for this 'reduced' pay?

Why would it be less to develop using open source than (possibly) reinventing proprietary wheels?

Simon Lucy
Monday, June 7, 2004


Don't you ever dream of being more than an employee? Do you enjoy the 9 to 5 grind?

If software is commercial, then you can write software and SELL it.

If software is open source, then you can write software, give it away, receive nothing in exchange, and you are supposed to earn from customization and tech support contracts.

To me, this seems a lot worse than the "write software, then sell it" model.

Also, not all software requires customization and tech support.

So.. from this "customization and tech support" deals only a handful of "stars", like Linus and RMS, earn money. The rest just write software for free, and are never rewarded financially for their work.

I am curious if there is any study about the real-world benefits that programmers gain by writing open source.

I am sure that for at least 70% of them, the benefits are extremely small, for another 25% there are some benefits like recognition, and only the rest of 5% have really benefited from writing open source.

Then who benefits? The users, who are glad to have the software for free, and the business people, who of course are extremely glad to save some money.

Why should we work for free for the businesses?

If the open source trend continues, software will become a commodity. Now, you have probably heard this several times.

But what you may not know is that in commodity markets survival is tough, profit rates are very small, and only the ones who are able to work very cheaply succeed.

You may say: I do this for a hobby. I enjoy writing software. Writing software is extremely rewarding in itself. If you think about it like this, it's fine, and I understand you, because I like writing software myself.

But you must also think of the future: what will happen to your salary if you and 100000 other programmers keep churning out free software?

My advice is: when you feel the need to release something as open source, don't. Just release it as $1 software, or make it free, but don't release the source.

Michael Schredder
Monday, June 7, 2004

Michael, You first start with

"Don't you ever dream of being more than an employee? Do you enjoy the 9 to 5 grind?"

Then you finish with

"But you must also think of the future: what will happen to your salary if you and 100000 other programmers keep churning out free software?"

I guess that in reality, you and almost everyone isn't worried with

"If software is commercial, then you can write software and SELL it."

Mainly because development is a difficult profession (hard to do a lot alone).

Someone has always to pay the bills. Somehow OS benefits the corporations. Perhaps they just want to lower their operating costs.

OS can be taken advantage of... Many ISVs exist because of OS.

IMHO. :)

Monday, June 7, 2004

Well, lets see.

I don't work 9-5, I work when I need to and for as long or as little as it takes. 

I don't base my billing on whether what I'm using, developing or fixing is open source or proprietary;  I base it on what needs to be done and how long I belive it will take, as well as how much I think I can get. 

Sometimes, if its using an open source product it will be cheaper because I won't have to do very much, but on the whole the consultancy needed to do the job balances that out.

The only limit I place on myself is that I won't modify GPL'd sources but that's only for the avoidance of doubt so that some way down the line I won't be penalising myself because I have code that's under GPL that I might also have under a proprietary licence (regardless as to who owns the copyright).

There is nothing to stop you selling Open Source software for as much as it can stand, whether you develop it or install it.  There is nothing to stop you selling anything, but you must be clear as to what you are selling.

This is an important point because developers especially seem not to understand it.  You are not selling source, you are selling a product and probably ancilliary services along with it.

The source is not the product, in much the same way as the blueprint is not the Shuttle.  The steps required from source to working product may be one or they may be fifty.  Each one of those steps can be billed for.

If you did but know it Open Source increases the earning power of developers that know how to (my god I'm going to use a marketing word), leverage it.

Just as those that understand the labyrinthine connections between MS products can leverage them and sell them though the actual value they add may be very small.

Right now, around 90% of what I do is using a proprietary licence, but that's because its for private clients and I let them have the source as part of that licence though I limit them to not producing derivative products.

This is because all of that development may be applicable to other clients at other times and I'll cheerfully resell some or all of it.  I could make it Open Source but we are using a framework as the basis which is not Open Source,

In six months time I might be working wholly in an Open Source environment again.

I really don't mind which.

Simon Lucy
Monday, June 7, 2004

On tides:

They always return.

Remember Canute.

The Brazilian example is instructive. In a world where information technology is becoming prerequisite, the neo-colonial attitude of western corporations utterly unacceptable to many. The manipulation of WTO rules on intellectual property and patents to suit (particularly, but not exclusively) American interests does not go unexamined and without comment.

In such circumstances the development of software libre has several advantages; it is politically acceptable, particularly in those countries whose agricultural exports are restricted by WTO rules; it provides a low cost  opportunity to participate as an equal in world class software development, thus maintaining and expanding the intellectual capital of the nation; and it facilitates the development of a software commons for the benefit of the many rather than the few.

If Albert's hypothesis is correct, which on a global basis I do not believe to be true, the only question is when will the tide return?

Monday, June 7, 2004

Is this one of those threads a Microsoft employee writes; as part of their Marketing campaign?

michael sica (
Monday, June 7, 2004

Gaius, out of curiosity - do you believe that lesser privileged nations should be able to steal the infrastructure necessary to join the global market? For example, tractors, fertilizer, irrigation equipment - just nick them from a neighboring country or hijack a freighter.

After all, the high price of modern farming implements is holding those nations back, right?

This isn't so much about open source as about the comment regarding the WTO's protection of intellectual property rights; though the defense of open source and the general tone of the post makes me suspect that you think underprivileged nations deserve an automatic step up by hook or by crook.

Now as a caveat - I fully support foreign aid and helping nations with lower standards of living improve conditions for all their citizens. I just don't think it is "owed" to them, nor do they have a right to take it.


Monday, June 7, 2004

> Is this one of those threads a Microsoft employee writes; as part of their Marketing campaign?

Albert is an MVP: he is given the MS kool-aid to drink, at conferences and so on, but isn't their employee.

Christopher Wells
Monday, June 7, 2004

Albert:  Get an editor.  You post was overly long and had an unpleasant tone that seems to have put people off.  I think you've managed to make a very persuasive argument FOR open source, because I certainly felt compelled to distance myself from you.

I know a number of developers who make their daily bread from open source software. I have been one for the last couple of years. These are jobs that wouldn't exist without open source software.  These are jobs that left a half dozen people maintain a service used by thousands, so they don't even need to think about email support issues.  Because the components are open sourced, the service could be modified to fit the exact needs of the organization, something not possible for commercial packages (and several commercial packages were examined).

Sure, open source software isn't the right solution for many people.  This is a know fact and self evident. Proprietary software is likewise not the right solution for many people. It hurts my productivity and imposes undesirable limits on my products.

Clay Dowling
Monday, June 7, 2004

"Is this one of those threads a Microsoft employee writes; as part of their Marketing campaign?"

I'm curious - do you honestly believe this?


Monday, June 7, 2004

Philo, either you or I are misunderstanding Gaius's point. I think he was explaining that Brazil needs to pursue free-as-in-freedom software. Nothing about stealing IP, except maybe not respecting most software patents.

Though on your point, I hear my country stole intellectual property from Europe when it was developing. (The US.) Does anyone have a source on this?

Tayssir John Gabbour
Monday, June 7, 2004

Open Source software doesn't have to mean software but possibly a service or technology?

Open source software is such a vague term now, is Jimmy's first C program that he puts on the web open source software, is linux open source software, is a linux mandrake distro open software, is http an open standard for the web traffic, is the original spec on html/javascript open source.

My point, open source will always be prevalent, whether it is the next set of internet standards or intenet technology or set open source libraries that exist on propietary as well as freesoftware platforms.

I think the company that utilizes both open source and commercial products will be successful, especially, depending on what your goals are.

For example, a person approaches you, he says, we need a custom desktop environment for government workers in China.  You could hack together a completely new operating system from scratch from the ground up.  The other option, you could use the open source technology, the linux kernel for your kernel, and if you want open source unix tools, you can install those as well.  And there may be aspects of your system that you want to develop yourself, there may be aspects that you purchase, or there may be libraries out there that are open source.

The Software Developer's task is a difficult one, there is no point in reinventing the wheel and implementing a technology that an open-source lab or university has been working on for decades and there is also no point in purchasing expensive libraries where open source libraries will do the job just fine.

And just as easily as there is open source C software, and open source Java software, I see room for open source .NET software as well.

Berlin Brown
Monday, June 7, 2004

Is it really fair to apply blanket conclusions reached for Linux to all of Open Source?

I don't run Linux, and the last significant *nix experience I had was 10 years ago. However, if Linux is anything like the *nix from back then, then I could agree with Albert's comments regarding Linux vs. MSWindows.

However, Open Source is way more than Linux. My favorite Java IDE is Eclipse, which is Open Source. I use Xerces, more open source, for XML processing. I'm investigating Hibernate, more open source, for object persistence, and I'm learning JBoss for J2EE hosting (more open source). The list can go on......

These are all top quality offerings (or at least seem to be), and it is unfair to compare them with an OpSys derived from something developed last millenium.

If Linux is Unix for open source, then I can readily see that Linux is as obscure and quirky as its ancestor. That doesn't mean that everything that comes from OS is obscure and quirky.

For that matter, proprietary software can be obscure and quirky.... ever delved into the stygian depths of MS COM?

Monday, June 7, 2004


Stealing farm implements (sp?) is a bit different that IP rules.

I think a better analogy is the patenting of genetic patterns found in seeds - whether those patterns are man-made, occur naturally, or somewhere in between.

That would be a better example of "open source" applications outside of software - in which case Gaius may have a point.

If your field of GMO free corn pollenates with your neighbor's field of Roundup-Ready corn, you are liable to Monsanto for royalties.

Does that make sense to anyone?

I think that Gaius's message is that Free Software provides developing nations a chance at infrastructure development.  And I think he is correct.

Monday, June 7, 2004

"Gaius, out of curiosity - do you believe that lesser privileged nations should be able to steal the infrastructure necessary to join the global market? For example, tractors, fertilizer, irrigation equipment - just nick them from a neighboring country or hijack a freighter."

Philo, do you believe in non sequiturs?

Using open source software is equivalent to stealing a tractor?  No wonder Microsoft seems more and more like a megalomaniacal, power mad enterprise, if this is indicative of the mind set of its employees.

Jim Rankin
Monday, June 7, 2004

Back on topic, an open source strategy, or at least the appearance of one, is now a requirement for any sizeable enterprise.  It's impossible to negotiate with Microsoft without one.  So investment in open source will necessarily continue.

As Mr. Kallal points out, though, this does not apply to small ventures, as they can't negotiate with Microsoft over price whatever they do.  Of course, they should be using Macs if they really want to save on administration costs.

Jim Rankin
Monday, June 7, 2004

Much too long a post Alber to reply to in its entirety but a couple of points.

---" What is incredible is that this kind of offering was not done sooner! I mean, it has been, what 6 or more years that we purchase pc’s with the OS installed?"------

Small Business Server which you are referring to has been availiable for six years or longer. The fact that you work all the time with small businesses and haven't realized suggests that it might not be quite the no-brainer you think. I'll explain why in a later post.

----" I mean, why in the year 2004 do we actually send a tech around to each pc to install a new version of office, or stuff like a new version of  Simply Accounting clients on each pc? "----

We don't; you might! I was cloning in 1999, when MS still didn't recommend it and you had to get your MCSE you had to read hundreds of pages about scripting for unattended installs. Since XP even MS has admitted that the normal way to install is to clone. This of course can be done with Linux as easlily as with Windows, though the much greater ease of a Linux distro install compared to a Windows/office/ other apps install makes the difference less clear. Keep the data on another partition.

----"Virtually all I talk to find a HUGE improvement in the quality of office xp (2002). What the hell happened to make this release SO MUCH better then office 2000?

It isn't. I upgraded on the laptop because of XML support in Access (and I will upgrade the work desktop to 2003 from 2000 because of XML in Word) but I do not see a great increase in stability compared to Office 2000 running on Windows 2000.

The point is that MS thoroughly tests one version of Office against its equivalent Windows version but only does desultory testing with the older version (and of course the testing has to be the other way round for later versions). So Office 2000 is not completely stable on XP, and was very unstable on Windows 98FE (win 98SE seems to have less problems).

'Nuff for now.

Stephen Jones
Monday, June 7, 2004

Regarding developing nations, it's highly likely they're being taken for a ride with the open source propaganda.

The most useful thing they could do for their economies are develop valuable IP including software products and companies. Instead they're farting around with free products, and will never develop strong industries.

In 10 years time, their flirtation with foss will be seen as yet another disaster in their struggling economies, while IBM, Tata and the rest suck up bigger and bigger consulting fees.

Call me Bill
Monday, June 7, 2004

"Instead they're farting around with free products, and will never develop strong industries."

"while IBM, Tata and the rest suck up bigger and bigger consulting fees."

so you predict that there will be a strong industry in developing countries based around providing support and consultancy advice for the opensource products that are in use?

Monday, June 7, 2004


Good to know.  Don't think this really refutes my point, though.    MS held out quite a while before they gave in.

Matt Conrad
Monday, June 7, 2004

No, there won't be a strong industry in developing countries providing support to open source products.

The companies doing the support will be foreign outsourcers, which won't help the locals or their industries at all. Revenue that could have gone to local software companies will instead go to foreign shareholders.

Call me Bill
Monday, June 7, 2004

"The companies doing the support will be foreign outsourcers,"

interesting position.

If local IT companies can offer support at competitive rates what is stopping them from doing the support?

Monday, June 7, 2004


You and I seem to be at loggerheads today, which I regret.

I'm afraid I simply don't follow your analogy with tractors or farm chemicals. Would I object to a poor country taking a _public domain_ design for a tractor building their own, or a _published_ formula for a generic pesticide or drug and manufacturing it, rather than buying it from Massey-Ferguson or ICI? No.

Do I think that they should be able to _steal_ the design or formula? No.

My point is that this forum always seems to focus on the western, strictly capitalist view that sets the _purpose_ of software development firmly as a means of making money. I think we are misguided when we think that the majority of the world shares this view. We, as programmers, see software as something to be bought and sold; in fact it is just a way of facilitating _other_ processes. Beyond immediate needs for clean water, a reliable food supply, transport and education, software is an essential element of economic development. Why on earth should a country _not_ use the leverage availed by open source and open systems?

The key issue, I think, is that the demand for software (particularly public sector, infrastructure software) far outstrips the ability of many countries to buy at western prices (cf. anti-retroviral drugs in Southern Africa) or presently to write. Open source offers a 'trading environment' that can be used to lever expenditure made in other fields, particularly in higher education and government.

I can't afford to buy a patient records system, but I can afford to share the cost by participating in an open source project. Such a decision can be made easily at a low level. The western alternative, of offsetting the cost of closed development followed by future sales, is impossible. There is limited capital available to undertake the development, and there is no governmental expertise in spinning off the organisation as a separate, trading unit.

Monday, June 7, 2004

Indeed - when I spend my time creating something, I generally expect to be paid for my time and effort, so that I may feed *my* family.

When a group of companies work together to create an open-source patient records system, what is usually happening is that the employees of the companies are being *paid* to write the code to create the system. The companies are paying the employees to produce a product.

This is all very capitalist.

You know, it's interesting that nations with a smaller GDP discourage paying for software - they are hurting themselves. Software is a wealth generation machine - the software author can write something once, then sell it many times. Purchasers are paying for value, so it's a valid product.

So in theory Elbonia could create a tech center - a T1 line, some steady power, and a few development boxes. Create a product people want to buy, sell it worldwide, and cash comes flowing into the nation's coffers.

Instead, the Elbonian government makes noise about free software and hamstrings themselves.

This is overly simplistic and off-the-cuff, but it's an interesting approach...


Monday, June 7, 2004

"Instead, the Elbonian government makes noise about free software and hamstrings themselves.

why are those two ideas mutually exclusive?

surely the Elbonian govt can use opensource software for all of its most common requirements.....word processor, operating system, spreadhseet etc etc and still encourage and promote its local developers to create and sell (overseas) propriety software?

Monday, June 7, 2004

actually the more I think about it, the more specious that argument is.

There is _nothing_ to say that the elbonian govt cannot mix and match as they see fit, and nothing to say that regardless of whether the elbonian govt is using an open source product to do its word processing; a local developer cannot still make their fortune by writing and selling a killer app.

help me out philo, what obvious point am I missing?

Monday, June 7, 2004

The problem is that open sourcers in those developing nations aren't advocating mix'n'match.

They think open source is good, proprietary is bad, and everything should be open source as quickly as possible.

That's what's wrong, and it's why those developing nations are shooting themselves in the foot. The foolish politicians now spruiking for IBM will end up paying massive outsourcing bills to shareholders in America, which is what they thought they were avoiding.

To the person who asked why local firms wouldn't do the support, it's because support has no valuable IP; it's just direct price competition, and big companies kill local firms on that. Look around. You buy from Dell, not XK's Corner Shop.

Call me Bill
Monday, June 7, 2004

"The problem is that open sourcers in those developing nations aren't advocating mix'n'match.

<g> surprisingly enough neither are the closed sourcers.

They dont need to, let those who are advocating open source go that way and let those who are advocating closed source go that way.

both industries will develop as far as they are able.

If the govt wants to insist on open source word processors fine, its not like any local company is likely to have created a decent competitor to word quickly enough for it to be used anyway.

"it's because support has no valuable IP; it's just direct price competition, and big companies kill local firms on that. Look around. You buy from Dell, not XK's Corner Shop."

that seems deliberately naive.  I buy hardware from apple, I buy it from dell and I buy it from local retailers depending on what I need.
But thats _hardware_ where the big boys have a lot of obvious advantages.

In computer support the big boys have few advantages and lots of disadvantages.  they have bigger overheads.

All else being equal, I can (and do) provide support for apple computers at a cheaper rate than some of the national companies here in New Zealand, I can think of no reason why that wouldn't also be possible in Elbonia.

can you perhaps enlighten me?

Monday, June 7, 2004

>I was cloning in 1999, when MS still didn't recommend it

Look, cloning is a laughable approach to managing and installing software on each desktop. For a full re-install, cloning is great (and, I even know some companies that can flag any pc on the network to do a full auto-install of  image (takes about 20 minutes right across their network on “next” logon time)).

However, I was not referring to imaging technologies. Imaging is NOT a realistic approach to managing your pc’s in your office. The bandwidth issues, and a whole bunch of other problems makes cloning a joke for “general” pc management.

Sure, imaging can save some costs.

We are talking about installing patches, or perhaps JUST the sales department gets the new version of Excel.

>and you had to get your MCSE you had to read hundreds of pages about scripting for unattended installs.

Yes. the above is rather my point. The above is like telling the first person who wrote text on a computer it was done with hand coded assembler. What the heck is any use of you having to write a bunch of code/scripts to install software? (did not everyone get my point here about usability at all?..and why MS is making such gains?).

So, you tell me that you can write scripts to do this? (how the hell does that help a company that does not have someone to write those scripts for?). I gladly accept your point that many companies WERE early adopters of technology that can reduce support costs (corporate edition of ghost is great example).

With SMS, you now get tools that allow a company to maintain, and manage these installs. This is all about usability.

You can check out the basic features of SMS here

>This of course can be done with Linux as easily as with Windows, though the much greater ease of a Linux distro install compared to a Windows/office/

Hum..really? I would certainly accept that writing scripts in both cases is easy..but that was NOT my point! You telling me that Linux has management tools like SMS? And further, what good are those Linux tools for the desktops?

Ok, then I am in the dark here.
Can you give me a nice link so anything like this for Linux?

>The point is that MS thoroughly tests one version of Office against its equivalent Windows version but only does desultory testing with the older version

Yes, so then why is office 97 so stable on a new xp box?....when office 2000 is not?

Office 97 is way more stable then office 2000? (hum?).

The fact is, that the code base for office 2000 was VERY much increased in size (likely too much in size!!). There was as a result a huge increase in bugs. Simply put, a better approach to managing the code, and how bugs etc would be fixed was needed.

The improvements were not just a issues of testing (I had lunch with the guy who runs the bank of servers where they do automated testing of office (it tries all of the features…one by one)……and the things he told me was incredible, but I don’t believe that office xp benefited that much from this automated testing). On the other hand, I do know some incredible (radical) changes they are making to those automated test severs (but I am under nda).

Albert D. Kallal
Edmonton, Alberta Canada

Albert D. Kallal
Monday, June 7, 2004

Hmm, cute.  You say you're trying to look at things "how they are", then proceed with page after page (!!) of pro-Microsoft, open-source-sucks ranting.

Your primary argument seems to be that Linux usability sucks.  Ok, it kind of has.  It's getting much better, though.  I could say that Microsoft's stability sucks.  That's also true, but they're getting better at that, too.  Two paths to the same goal.

You also go on for a while about how Microsoft's automated crash reporting means it's going to have more reliable software.  More reliable than what?  If you mean that Win2k is more reliable than WinNT4, ok, true, but it probably would be, anyway; this has nothing to do with Linux.  If you mean that it's going to be more reliable than the corresponding Linux apps, you should first notice that many Linux apps already have automatic crash reporting.

Sounds like somebody's bitter.  (Couldn't get Redhat to install?  Boss yelled at you because you picked Linux?)

Tuesday, June 8, 2004

Wally: Get with the program.  Open-source *is* dying.  This isn't news.  It's been dying for 15 or 20 years now...

With apologies to David Thornley: "Doesn't look any deader than usual to me."

Tuesday, June 8, 2004

*  Recent Topics

*  Fog Creek Home