Re: Accelerated legacy code



On 2009-12-06 18:01:30 -0600, "Pete Dashwood" <dashwood@xxxxxxxxxxxxxxxxxxxxxxxxx> said:

Cydrome Leader wrote:
Pete Dashwood <dashwood@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
Howard Brazee wrote:
We tend to associate "legacy" with "mainframe". But it ain't
necessarily so.

From my limited exposure, the large new systems I have seen have
lots of code which doesn't fit new standards, but is kept around
because redesign and replacement is much more difficult when
anything can touch anything.

I'm calling it "accelerated legacy code".

I used to think "legacy" only applied to mainframe until I was given
a task to "keep legacy running while we develop and implement a new
solution" for a major Utility in the U.K.

heck, there can be legacy code and applications from even a few years
ago.

One example is a CRM/ticket system that only works with IE6. Users
can't even access it as the desktop build is beyond IE6. It simply
doesn't work, even with IE7 or IE8 in legacy mode. It doesn't work
with other browsers at all.


That's a pretty bad situation and a lesson to us all. I think most
responsible web developers check for at least SOME fundamental cross Browser
capability, or issue a disclaimer on their sites. With modern tools it isn't
TOO hard to check things across Browsers (although it is definitely extra
work)... 5 years ago such tools did exist (MacroMedia have had this
capability in Dreamweaver for more than 5 years, just as one example).

It's looking like some windows 98 machines will have to be kept on
hand to use the thing, and it the data on it has to be accessible for
X more years, even if nobody knows how to run it and it's on hardware
with no more support. It has to be less than 5 years old too.

I have to wonder why anyone would take delivery of such a system. I guess
they saw functionality they needed and decided the tech stuff didn't
matter...

I would have thought it would be pretty easy to replace a CRM system; there
are dozens of very good packages around that provide that functionality.

Pete.

In general, I would find myself questioning the facts in this case. Five years ago, in 2004, XP
would have been the standard build. Windows 98 was at that point, around six years old.

And even if the facts supported the idea of the necessity of keeping Win98 machines around,
in 2009, those "machines" would be virtual machines running on top of Windows Server 2003,
Windows Server 2008, Linux, or MacOS.

-Paul

.



Relevant Pages

  • Re: The destruction of the C99 standard
    ... In the end we had to perform several rounds of code-size optimisations ... shame if the standards encourage laziness in this sort of area. ... little demand for a feature, but that the feature is impossible or ... variety of machines, many of them *much* smaller than 32K. ...
    (comp.std.c)
  • Re: Pinball Video Repository?
    ... Very important to decide on standards for resolution, ... today and it's absolutely loaded with all sorts of machines and brand ... density improved), once we start talking video, it goes really fast, ...
    (rec.games.pinball)
  • Re: Question on strncmp / strnicmp use
    ... You mention increment of pointers and comparison, ... There are machines where strlen can be implemented as a single ... Standards seem to have a life of 10+ years while hardware development ...
    (comp.lang.c)
  • Re: Anyone here tried KBasic? [OT]
    ... It was a hugely important issue for the bigger businesses as well as the more thoughtful smaller ones. ... I'm sure that standards played a big part in the market dynamics of the time which is part of what I meant by the "suits" at IBM and Microsoft. ... They were replaced with Amstrad machines that were not PC clones but didn't have the looks of the Mac either. ...
    (microsoft.public.vb.general.discussion)
  • Re: A Brief Look at History
    ... standards are a necessity to produce maintainable code. ... A large part of good management is detecting the need and directing help where it's needed. ... and you have to accommodate a diversity of standards which still contain holes. ... That's the kind of collaboration that produces shared code bases. ...
    (comp.lang.forth)