Sign in / up

back to article Old apps must die when you migrate to the cloud

How many applications is too many? In March, Capgemini issued its 2011 Application Landscape report, which surveyed almost 100 companies application portfolios. It found that 60 per cent of enterprise respondents had more applications than they needed. In large and enterprise-class firms, a bigger proportion of people felt …

COMMENTS

House rules Send corrections
This topic is closed for new posts.
  1. The BigYin

    If the apps were open source...

    ...there'd be less need for the teams to re-invent the wheel.

    Your company's worth is in the business process and rules that the programs follow (which should get loaded from config files) not in the code you write.

    0 0
  2. Anonymous Coward
    Boffin

    Never going to stop

    The problem is quite clearly highlighted in the article, you get the applications down, get people retrained, have the system reduced to the bare bones, then some one decides to buy another company.... Rarely are the people in IT asked about the costs a new acquisition will incur to migrate the systems, in addition to financial audits there should probably be a requirement for a migration audit prior to purchase only then would a company not be doomed to repeat the mistakes of the past.

    0 0
  3. Anonymous Coward
    Anonymous Coward

    In other words: The hyped cloud can't run half your applications, or at least the salesmen failed to mention the costs of porting and migrating. But fear not: it's all the customer's fault for wanting them in the first place. If you just rewrite them and re-train your staff, you can get most of the functionality back within a year or two.

    "In large and enterprise-class firms, a bigger proportion of people felt that there were far too many applications." What a surprise: the accountants don't understand why the engineers need three different finite element codes, and the engineers think money is wasted on SAP.

    2 0
    1. JimC
      Reply Icon
      Holmes

      AC/What a Suprise

      > the accountants don't understand why the engineers need three different finite elemnt codes...

      'exactly so. The people saying this are rarely those whose work relies on the despised extra applications, and whose life will be made much more awkward by a half-arsed semi functional replacement...

      And guess what: if management were to budget the work needed to tweak the new apps to do the job properly the whole business case for replacement would fall in a heap on the floor...

      Too often we forget that IT is there to help the users provide a service to the customers rather than be an end in itself.

      1 1
  4. Anonymous Coward
    Anonymous Coward

    Crap management

    IMHO, the 'bloat' is due to crap management. I've worked in places where we had 30+ year old stuff doing relatively trivial jobs, which should just have been refactored and re-engineered. The alternative would be to go back to the customer and convince them that they don't need/want the application any more (in a lot of cases, I'll bet they just use the app because "they always have" and because "my boss asked me to email the report to him every week", and not because it's actually doing anything useful for them). I'll bet that those customers weren't getting charged anything like what that application actually cost to run either - had they been, they probably would have worked out they didn't need it after all.

    As a manager, it's far easier to create new stuff and get kudos for how great you are than it is to work damn hard getting people unhooked from the old stuff. The latter requires time, dedication and an ability to work without any praise or recognition. Who's going to want to do that?

    Strong IT management isn't just about supporting the latest iProduct or Cloud, it's about cost control, which includes legacy removal. Unfortunately, corporate culture usually rewards doing easier, good looking stuff, and not doing the hard, boring stuff. I'm not sure "the cloud" is going to do anything positive or negative about that.

    2 0
    1. Peter H. Coffin
      Reply Icon

      Save that keeping the old is often cheaper than retooling process. How do you think giant mainframes and COBOL have lasted this long? They work, they work the same way that they did last year and five years ago, and do so reliably enough that a Director of IT can swear in court that these results are exactly as reliable as those from a decade ago. There's a reason that unhooking people from old stuff is so hard: they depend on it.

      0 0
  5. Figgus
    Thumb Up

    "As a manager, it's far easier to create new stuff and get kudos for how great you are than it is to work damn hard getting people unhooked from the old stuff."

    +1, just for this single line. Repeated because it cannot be said enough!

    3 0
  6. Peter H. Coffin

    Bespoke applications

    So the recommendation is to set aside the custom-built applications, honed over decades to do EXACTLY what the business wants and needs, using processes that are difficult if not impossible to replicate in anything other than another custom application, which (if changed) will provide different results are require readjusting or retooling ALL the information handling and decision making downstream, if one doesn't get a new custom app written that works with The Cloud?

    Good. That's my field.

    0 0
This topic is closed for new posts.

Other stories you might like