Pure Play Maintenance Costs Consuming IT Budgets?

ContractualObligatio asks: "The Register asks the question of whether businesses are at risk of having no budget to develop code, from maintaining too many 'pure play' applications. What has the experience been among Slashdot readers? Are people spending too much time maintaining code and integrations because the business is adopting too many applications? Do IT teams have the time and money to actually improve and innovate the way their companies do business?"

Comments

pure play?

[tepples]

The article doesn't define "pure play" applications. Is this a common term of art in IT? Google shows "pure play" defined on investing sites as a firm that concentrates on one type of product.

I'm going to guess here...

[khasim]

I'm guessing that this means an application that does only one thing. Seemingly with a limited API or hooks to link it with other apps. From TFA:

But in practice, the level to which any pure play application integrates with an infrastructure management suite can be a significant variable, and there is always the possibility that adding something like lifecycle management means adding yet more pure play applications to the maintenance list.

So, kind of a limited version of the old *nix ethic. Do one thing, but do it well.

So their point seems to be that you want to evaluate the ability to hook your apps together when you consider which ones to purchase.

Oh, and remember that SOA is a cool buzzword.

"Pure play"?

[illuminatedwax]

Yes, I would prefer to have "virtual work" or "hybrid play" in my IT office! See I can make up words too!

Re:Too much to do to get work done? What?

[afidel]

Only true if you can make an application design that can scale arbitrarily with the business and can meet all current and future business needs. Since that's impossible there will always be some level of maintenance and replacement of software systems. Then there's the hardware platform on which the application runs, and unless it's something basic like OS/360 nee Z/OS then you might have to redisign the app every so often to keep up with the changing OS platform. The goal of IT is to meet the businesses needs at the minimal cost necessary to maintain that position and meet ongoing business growth.

Re:Too much to do to get work done? What?

[eschasi]

You write it once, you write it well, and it works, and keeps working.

...giggle...

...snort...

hahahahahahahahHAHAHAHAHAHAHAHAHAHAHhahahahaha . . .

Go on, pull the other one.

OK, ok, I'll be serious for a second. What about:

• OS patches that break things
• new feature requests from users
• new browsers that have to be accomodated
• the code just being slapped together in the first place (yeah, I know you said you write it well. Given the circumstances under which most non-programming IT shops build apps, it's rarely written well.
• Applications that interact with it change
• etc, etc, etc.

Let's face it, the circumstances you describe are damned rare.

text is too large - OMG!

[Nataku564]

Because it affects a good chunk of us at work. At least, from my experience, it does.

I work at a pretty decent sized hedge fund. We have been growing since before I arrived, and continue to do so at what I feel is a fairly rapid pace. In order to seem "current" and keep up with the market, our trading desk has been expanding into unholy amounts of new investment strategies. Lots of fun OTC derivatives, energy markets, weather, pollution trading, etc.

This is all well and good on the face of it, but in order to support these new strategies as quickly as they claim they need to, we need to purchase outside software instead of developing it in house. So a good portion of the IT staff (myself included) has been entirely devoted to figuring out how to integrate these new applications with our existing structure. Basically - theres no time for new code, and its frustrating most individuals involved, since we kind of like to write code, and not attend multi-day training sessions on how to support some lame half-engineered application that does all its work in stored procs from a single threaded windows form based server. No joke on that last one either, that app bites hard - yay for me being in the department in charge of that monstrosity.

So, its a common problem - or, at least I think it is. I guess this article is supposed to foster discussion on that idea. Of course, its not actually a problem with the IT staff, but we still have to deal with it. The staff in my area alone has doubled in about a year's time in order to try and keep up.

Re:One step further

[schematix]

while(time < fiveoclock)

Ahem. Not to be too anal about your code, but this would certainly cause some issues (most notably with disgruntled workers at the office at odd hours.) Perhaps you should add some additional logic such as (time > eightam), (time <> lunchtime), etc.

Oh yes

[mwvdlee]

As somebody who has worked on a multi-million dollar project that would bring in about $20,000 each year for a product that would be phased out in about 10 years and could have just as easily been replaced by a paper form... yes, too many "pure play" applications.