Fog Creek Software
Discussion Board




Phil Greenspun

Why would a guy who preach: Get the code out. Get it out. Don't sit there and waste time doing design. Blow his company out of the water by spending massive amount of resources building this super duper site building framework in a language nobody at his company knew well and had all kind of performance and evolution problems?

I'm just picking on Phil cuz everybody knows him. However, this was a VERY common problem at your typical dotgone company. Aeron chairs are expensive but they're nothing compared to high salaried inexperienced programmers sitting around creating very complicated systems that only developers with 10+ years of experience can do well.

FYI: I do not consider you experienced until you have at least 3+ years under your belt..and even then, you're probably not a hyper-productive programmer yet because you're still letting everyday "distractions" affect your concentration.

Humbug
Saturday, February 28, 2004

I'm not 100% sure what you're talking about -- sounds like you're griping about the Java port of ACS -- but that's not what blew arsDigita out of the water. What blew arsDigita out of the water is that the market for Internet consulting services collapsed dramatically -- many companies I talked to in those days saw a collapse of 90% of their business over the course of one month; by January 2001 literally every signifacnt US company with money to spend had a moratorium on any consulting projects and had kicked out the consultants they had.

What's worse, the consulting firms didn't know it was happening while it happened. They just thought that the sales cycles were taking a little longer, or, in the case of arsDigita, they thought it was the VC's fault that the customers were disappearing. In the meantime they went from having 10 clients for every programmer to having 10 programmers for every client. They didn't want to lay off programmers, or even think to do that, because demand had been so high over the previous two years that the only limit to the amount of money you could make was how many programmers you had. Everything about these businesses was oriented to hiring as many programmers as possible so as to be able to take all those clients who were lined up outside their offices with wheelbarrows of crisp $2000 bills. (For example, a component of the Fog Creek business plan was that by providing better working conditions and private offices for programmers, we would better be able to recruit the programmers that attract the cash wheelbarrows.)

It was impossible to predict that the market would collapse so dramatically and so suddenly, and almost every consulting firm in that market (Scient, Sapient, Viant, Razorfish, IXL, arsDigita) collapsed because they were still paying overhead and salaries but without any revenues to support it. At Fog Creek we also saw our consulting businesss disappear in those months. We survived thanks to license revenue from our software products... in many ways we were extremely fortunately that our business plan was to be "arsDigita with commercial, not Open Source software" or we would be gone now, too. You can read about it in my article "Spring in Cambridge" which was written after the market for consulting had cratered but before anyone realized that it had.

Joel Spolsky
Fog Creek Software
Saturday, February 28, 2004

PS it's "Philip", never "Phil"

Joel Spolsky
Fog Creek Software
Saturday, February 28, 2004

Joel:  "Spring in Cambridge" ???

I'm very fond of your writings, but there are quite a few of them.  PLEASE put real links in your answers so we can navigate directly to the page.  Especially as your 'search' capability doesn't seem to work very well.

Thanks.

AllanL5
Saturday, February 28, 2004

It's not a specific grip about Java. It can be applied to any new technology.

Anyway, so the reason why they had tons of programmers working on their "framework" was because they didn't have any more consulting jobs for them to do? Not because they over-hired for the purpose of building the framework?

I see this strategy as being quite common among many consulting companies. Note: The focus here is on the framework, re-usable tools etc. I'm even aware of one local consulting company that spent considerable amount of money on a visual tool so that a user can model process, have it generate code and not require programmers. Silver bullet anyone? Sure, they may have placed a few of their "on the beach" developers on that project but they also hired very expensive (but inexperienced) people for the sole purpose of building their super tool.

I think there's too much being blamed on bad market conditions and not enough on bad management "engineering" decisions.

Humbug
Saturday, February 28, 2004

I worked at arsdigita. The consulting market collapsed, the company overhired (by 200 people!), we spent too much money on stupid crap (haunted mansion on cape cod, a leased ferrari), we gave huge bonuses to the wrong people, the founder sued the company for its cash reserves,  blah blah blah, and so on, and so forth.

Any stereotypical dot-com mistake that could have been made, was made.  The "management" was to blame, the market conditions were to blame, the programmers were to blame, and the boss was to blame. Even if nobody was hired to work on the "product", it isn't clear the company would have survived the market collapse. Even if the market collapse never happened, it isn't clear that the company would have survived the chaotic mess it had become.

 arsdigitan
Saturday, February 28, 2004

spring in cambridge:  http://www.joelonsoftware.com/articles/fog0000000021.html

The Ted
Saturday, February 28, 2004

...don't forget DeathMarchFirst.

ok, ok
Saturday, February 28, 2004

You so nailed this, Joel.  While arsDigita may have had their own internal problems, it's pretty obvious that what caused all the other consulting companies to bankrupt at the same time probably played the most into why arsDigita fell apart.  For this reason, I have never been particularly enamored of arsDigita or its philosophies.  After all, a business can only have a social agenda if it makes money and if it has a workable business model, and it's not clear that arsDigita could sustain either.

Richard Kuo
Sunday, February 29, 2004

"It was impossible to predict that the market would collapse so dramatically and so suddenly..."

I disagree.  In fact my memory of the year or so leading up to the crash was me and a lot of other developers amazed at what was going on and wondering when the speculative bubble was going to burst and kind of delighting in anticipation of all these paper millionaires (who thought they were so smart because they were paper millionaires) realizing that they were suddenly normal people again.

It was quite apparent that once the bubble burst, the small companies who did nothing but consult to dotcoms would quickly be in the crapper.

To me this is one of the oddest things about the bubble, that everyone acts as if it was a complete surprise that it burst.  Sure the exact time was a surprise but the papers and journals were filled with articles comparing it to past bubbles.  The Chairman of the Federal reserve kept telling us in different words that it was a speculative bubble.

I think what you really mean is that there was no way for people caught up in the mania to realize the market could crap out so quickly and so dramatically.

name withheld out of cowardice
Monday, March 01, 2004

What I found amazing was the bubble lasted as long as it did. The guy in the deli in our building got pre-ipo stock of some internet company because he made sandwiches for the CEO. Once he started giving out stock advice, it was obvious that it was way overblown. It was almost a year before it finally crashed.

The deli guy sold out and retired so I guess he was pretty smart after all. The company went bust, or is just hanging on.

pdq
Monday, March 01, 2004

I am one of the few unremarkable people who have taken the time to read a significant part of the ACS source, Humbug, I don't know where you get these opinions from, but they don't appear to be very well based. If you just want to stir up some emotions---do it at RedHat, they are the current owner of this "piece of failed technology".

Li-fan Chen
Wednesday, March 03, 2004

My facts came from a developer who actually spent time working on that code base. Again, ArsDigita was just an obvious and easy example to use. I can apply it anywhere else. I have friends working at other consulting companies doing all kinds of "silver bullet" work with technologies that they are not that familiar with. I'm sure even today there's some people grunting it out in .NET now that Java is "old school". Please understand, this isn't an attack on Java or Ars Digital. I could probably use C++ and some other company but then my development memory just doesn't travel back that far.

Humbug
Thursday, March 04, 2004

Actually, scratch my previous post. My original point wasn't about companies making the same mistake over and over again. It was about Philip allowing his company to spend time doing framework stuff. I could swear that his book writes against these things. Maybe I should go back and read it again?

Humbug
Thursday, March 04, 2004

I read Philip's narrative of the demise of arsDigita. It is quite enlightening. IIRC, he was against the "grand re-implementation" actually.

Really, arsDigita is a beautiful example of how fundamentally corrupt the VC arena was during the bubble. Once arsDigita accepted VC funding, Philip started to lose control of the company. Even though he held the majority of the stock, the VC's finagled control of the company, and they started pursuing the "fashionable" trends in software, and spending went crazy.

When clients started disappearing, but spending stayed high, Philip started trying to regain control of the company, couldn't, and ultimately ended up suing the company. It's weird when the majority stock holder can't get the CEO to resign.

The ultimate lesson seems to be, get a DAMN GOOD LAWYER, and don't trust one associated in anyway with other people you deal with.

Brad Siemssen
Saturday, March 06, 2004

Dear Brad,
                  There are a hell of a lot of people who claim that it was Greenspun's stupidity and not the VC's that sunk the company. Just look at the crazy spending he intitiated.

                    And I find it interesting that not even Joel expected the crash. Now if all these geniuses had bothered to read their John Dvorak they would have known about it eignteen months before hand. And if they didn't believe internet commerce was going to crash, they could at least followed his advice and bought a truck. Trucks are a lot easier to get rid of then pool tables, gold plated smurf balls, and Aeron chairs.

Stephen Jones
Sunday, March 07, 2004

I think it was John Dvorak who said that Windows 3.0 would be rapidly dismissed by most people, and they would go back to using DOS.

So apparently there are no such things as (working) crystal spheres.

Dario Vasconcelos
Monday, March 08, 2004

Dvorak got that one half-right:Windows 3.0 was rapidly dismissed by most people; they went on to Windows 3.1 :)

With regard to the dotcom boom he was spot on, as he was with regard to rubbish such as push technology and internet portals.

I rather suspect Dvorak is right as long as what he wirites about has nothing to do with computers :)

Stephen Jones
Wednesday, March 10, 2004

*  Recent Topics

*  Fog Creek Home