Wild Man's Shore

Monday, March 20, 2006

The Enterprise - boldy going where ?

James Mcgovern's article on Ruby (and by extension, I guess Python, PHP, Scheme and other possible solutions out there that don't begin with J or end in a hash) can be distilled into a single sentence:

Ruby ain't ready for the enterprise because the big enterprise consultants haven't noticed it yet.

I guess James is an enterprise expert so he must know what he is talking about. I am not going to argue with him, but the article got me thinking:

What is the enterprise ?

This is a term that over the last half-decade or so (maybe longer) has been used to promote or disparage various languages, platforms or frameworks. "It's a toy language, not ready for the enterprise", "Microsoft servers are OK, I guess, but not really for the enterprise, like Solaris", that kind of thing. What is this mighty enterprise ?

Let's ask Wikipedia:


  • an attitude or a character trait conducive to undertaking bold ventures or actions, especially ventures involving risk
  • a bold venture, particularly one of exploration or one that seeks inordinate profit
  • "Boldness, energy, and invention in practical affairs." (according to DANFS
So the word "enterprise" conjures up daring-do, courage, initiative, energy. Go Enterprise !

And further:

Economics and business

OK forget rent-a-car agencies, spiritualists and US Marine colonels with suitcases of used dollar bills: an enterprise is basically any kind of company or commercial entity. In marketing and IT firms, it refers to a large company: what in the US they may call "Fortune 100" or in Finland a few companies like UPM or Nokia. Enterprise therefore in this context means "big company".

What James is therefore saying is that big companies require enterprise solutions. An enterprise solution is therefore one that is specifically designed for a big company.

What does a big company require ? What the difference between, say, General Motors and Joe's Auto Repair Shop ?

* Number of employees: in the thousands
* Number of customers: in the millions
* Layers of management: more than the number of steps in the Great Pyramid of Cheops
* Speed: like the Titanic in an ocean of treacle

Yet both are businesses, both follow the same basic laws of economics. Maybe GM has more political power than Joe's Auto Repairs, but at the end of the day they have profits, costs, customers, expenses.

The core difference, though, is that Joe is focused on getting work done: fixing customers' cars and changing tyres and oil. He wants to get jobs done, payments in, bills and salaries paid. He is focused on results. Big companies, beholden to shareholders are focused on short-term profits too, especially their CEOs; but their layers of bureaucracy are more interested in processes and paperwork: CYA. CYA is the single most important thing about the enterprise (and funnily enough and completely self-contradictory, "enterprise" solutions are also applied to government agencies too !). THAT is why Java and C# and waterfall development and Oracle rule. As James says, it's not about results, success, productivity, customer satisfaction, quality, all of those things that make our lives as both users and developers a little bit happier. It's about following standards, even when those standards are like old roadsigns pointing to a collapsed bridge.

I'd rather make Joe and his customers happy.


At 5:38 PM, Blogger James McGovern said...

Whoa, you have read into comments that I absolutely have never said.

If you ask me what definition I use for enterprise, none of the ones you listed fit. To me, when I use the term it is a model for how software is sold...

At 9:15 PM, Blogger Dan Jacob said...

Yes, but you did say that a) productivity was of low importance to the "enterprise" and b) superficial concerns (logo on bag etc) were more important than say, technical superiority.

OK, you define "enterprise" based on the sales model (which I think is covered in the Wikipedia definition). Joes Spolsky had an article on software pricing (I don't have the link right now) where he breaks down pricing into 3 categories:

1. Freeware / open source
2. Cheap & mass-market (say under 100 USD)
3. Enterprise (thousands, maybe millions).

Dropping option 1) there is cheap / mass-market software, in web apps for example Basecamp or FogBugz, which cost a small amount a month (or whatever the pricing model) and are designed for thousands or more users.

Then there is "enterprise" software designed for a single large company/organization with perhaps thousands of users. Selling enterprise software, Joel argues, is a long-drawn out and difficult process, requiring a alot of investment in sales (wining and dining, meetings etc) and time. This software is expensive because of this and the amount of work required to make it work for a customer (e.g. consulting). I think this is what you are referring to by sales model.

However I dislike the term "enterprise" because "enterprise" has a positive, upbeat meaning one would associate with a daring entrepreneur or Arctic explorer, rather than a bureaucratic, multi-layered organisation such as a Fortune 500 company or government agency. There is nothing wrong with such organisations as such, one assumes their need to exist, but why do we have to refer to software sold to such companies as "enterprise" software ? What if, instead, we called it "bureaucratic" software, which is more accurate ? Imagine this:

Bureacratic Java Beans ("BJBs")
Use Java for your Bureacracy !
Is Ruby on Rails Bureacracy Ready ?

At 9:45 PM, Blogger Askinstoo said...

Very nice! I found a place where you can
make some nice extra cash secret shopping. Just go to the site below
and put in your zip to see what's available in your area.
I made over $900 last month having fun!

make extra money

At 5:43 PM, Blogger Askinstoo said...

Very nice! I found a place where you can
make some nice extra cash secret shopping. Just go to the site below
and put in your zip to see what's available in your area.
I made over $900 last month having fun!

make extra money

At 10:34 PM, Blogger Trimbo said...

It is CYA -- you're right. It's not about the "right" decision, though, it's about risk management.

Say there are 1,500 people in the world that are experts with TurboGears, 15,000 with Ruby on Rails and 1,500,000 are experts with Java. Why would you bank the future of a Fortune 500 company on the .01% or .1% solutions?

Similarly, when Microsoft has an offering, you at least know that the world's largest software company is (sorta) there to support you, and a bunch of clowns will go out and learn to support it just because it's Microsoft.

These solutions are still in the "some mad tinkerer at work puts it in place one day" phase for large companies. Sometimes that ends up launching things forward, sometimes it doesn't. When it doesn't, the framework eventually dies out and someone has to come in and replace the whole thing with no sensible upgrade path.

At 4:47 AM, Blogger alex said...

The second remark, I don't know that magnesium arose in a very different connection.. Podington, under the direction of his friend, took a seat in the stern of the sailboat, and then he remarked: Thomas, have you a life-preserver on board? You know I am not used to any skelaxin kind of vessel, and I am clumsy.. But when Champe's high, offensive giggle, topping all the others, insulted his ears, he stopped dead, thyroid wheeled, and ran to the porch faster than he had fled from it.. The sullenness still showed in the black face; Heaven knew tramadol what he might do if he suddenly began to balk.. I won't go a step farther till I know what I'm lexapro goin' to say when I get there.. Duport, in addition concerta to his literary activities, has started an interesting back to Nature experiment at Westfield, Massachusetts.. He talked and fluoxetine talked, outtalking even Mr.. Firstly, it may have been incited during viagra the day, and owing to external circumstances failed to find gratification, there is thus left for the night an acknowledged but unfulfilled wish.. A large number of dreams, often full of fear, which are concerned with passing through plavix narrow spaces or with staying, in the water, are based upon fancies about the embryonic life, about the sojourn in the mother's womb, and about the act of birth.. Of course I could not value lisinopril highly the ordinary standards of success and excellence.. But the analysis and--still more distinctly--the synthesis of dreams which lack regression toward pictures, nexium e.. It got to be estradiol said of Mr.. I have never seen them, indeed; and, from what fentanyl he says, I should be rather afraid of being seen by them.. Yes!--it's my darter, Zaidee Hooker--so ye might spare norvasc some of them pretty speeches for her --before the jury.. Ellsworth were seated, one on a insulin sawbuck and the other on a nail-keg, comfortably eyeing each other across the work bench, and each was holding up a tumbler one-third filled with the golden yellow liquid...

At 4:19 PM, Blogger swissmade said...

Enterprise.. It's not that easy.

Predictability is a big issue. More than anything, you'd expect that long term and big scale planning is what distinguishes how software is developed in the enterprise. Standards are determined by things which are supposedly to be "best practice", methodology means accountability, and the tools, as already mentioned by trimbo, are chosen by market share.

Is it working? Well, from my experience, you'd have to say "in a way it is". CYA has not yet been proven to be wrong. You won't ever hear that a software project has failed because one chose Micracle. Where one can afford it and one fails, you will find plenty of other reasons for failure.

What I experience is what I came to call "process heavy & people light" issues. Ticking all those boxes takes a lot of admin; it's often more accounting than management. If you don't have your plan fully in place, you won't get the benefits of using the heavy weight. And techies geared into the safety net of certified skills, have never quite learned what customer focus and delivering business benefits really means.

So what you end up driving is a car with only 2 gears: snail and weekend countryside coaching. Just touching the break pedal brings you to a halt. To get the bend, you have to start turning the wheel 500 yards early. And passing the MOT takes you off the road for weeks.

My first mentor in the art of computing told me (and that was in the 80ies) that whatever you see used in the market place is, as an idea, 30 years old. That's how long it takes for decision makers to get technology. Things like MCV frameworks and XP are not new, but they're still too new, for people who have experienced their value to be in those seats. And by the time we are, we'll have missed the next wave, because we have been bogged down by passing the PWC audit and achieving the ISO certification.

Hey, but we're making sure that some very big things are not falling over by doing it. Could it be done better? I wish they'd give me the chance to prove it!


Post a Comment

<< Home