[Info-vax] Any stronger versions of the LMF planned ?, was: Re: LMF Licence Generator Code

[Bill Gunshannon]

On 8/8/21 10:37 AM, Arne Vajhøj wrote:
> On 8/7/2021 8:24 PM, Bill Gunshannon wrote:
>> On 8/7/21 7:40 PM, Arne Vajhøj wrote:
>>> On 8/7/2021 6:42 PM, Bill Gunshannon wrote:
>>>> On 8/7/21 5:56 PM, Arne Vajhøj wrote:
>>>>> On 8/7/2021 7:59 AM, Bill Gunshannon wrote:
>>>>>>                                                       Go read some
>>>>>> of the stuff on LinkedIn about "Legacy Systems".  Not specifically
>>>>>> about VMS but the attitude is even if it still does the job if it
>>>>>> is old (ie. COBOL) it is bad and a problem.
>>>>>
>>>>> Being old is not a problem in itself.
>>>>
>>>> Being old is never a problem in itself.  I'm old and regularly
>>>> compete with people less than half my age, successfully.
>>>>
>>>>> It becomes a problem if:
>>>>> - it is out of support
>>>>
>>>> Lack of support for one part of an IS should not be a reason to
>>>> abandon it in its entirety.
>>>
>>> If that part cannot be replaced: yes it is.
>>
>> If your running a VAX then it might be a problem.  But, believe it
>> or not, most legacy systems are not running on old or non-existant
>> hardware.  VMS being the main exception.  :-)
> 
> Things goes out of support all the time.
> 
> CPU architectures (ISA).
> 
> OS. Either completely or on specific CPU architecture.
> 
> Database servers, web servers, application servers,
> message queue servers, cache servers etc.. Either completely
> or on specific OS or on specific OS and CPU architecture
> combination.
> 
> Libraries.
> 
> Shrinking OS are obviously harder hit than growing OS, but
> it is far from a VMS specific problem.

But, replacements are available for all of the things you
mentioned above.  Unless it's more about religion than getting
the job done.

> 
>>> And even if that part can be replaced then the question is at what
>>> cost compared top the replacement. And it also raises the question
>>> about whether other parts will go out of support soon.
>>
>> As has been stated here numerous times in the past, unless you are
>> running custom hardware and doing things like device control this
>> is not likely to be a problem.
> 
> Of course it is a problem.
> 
> Remember what happened when Oracle announced that they would
> drop support for VMS in Oracle DB client library? Not only was it a
> problem for those using Oracle DB, but it also worried those
> using Oracle Rdb a bit.

See comment above.

> 
>>>>> - it is hard to find people with skills
>>>>
>>>> That is a fixable problem.
>>>>
>>>> https://edscoop.com/college-legacy-programming-langauges-grant-bill/
>>>
>>> That is a good proposal.
>>>
>>> But do you expect serious companies to base their future on that
>>> such a bill get approved, 
>>
>> I  have little doubt that it will be approved.  Financially it is a
>> totally non-apparent bump in the budget.
>>
>>>                            that funding will continue in the future
>>
>> That will depend on whether or not academia decides to swallow their
>> pride and get behind the idea.  I am doing what I can to try and help
>> it, but for totally non-technical reasons it is going to be a hard sell.
>>
>>> and that students will be interested?
>>
>> I had students interested in legacy systems when I still worked at
>> the University even with members of the faculty attacking much of
>> what I was selling.
> 
> Bottom line: lots of hope but nothing sure.

The one thing that is sure is what Cartwright said, if we do nothing
the problem will get worse.  Using only one facet, COBOL, I have seen
the number of jobs looking for COBOL programmers go up by orders of
magnitude over the last year or so.  It was a problem before but I
think some of the side effects of the pandemic have exacerbated it
resulting in a recent major change.

> 
> Most businesses will prefer a technology where they know
> they can get people over a technology where they hope
> they can get people.

Funny, none of this has eaten into IBM's business at all, at least as
far as I have seen.  And when you see the word "legacy" in the mentioned
article, they mostly mean IBM.  Sadly, VMS isn't even oin the radar.

> 
>>>>> - it is expensive to maintain
>>>>
>>>> In the case of legacy systems expense is more objective than
>>>> subjective.  A little research will show how the majority of
>>>> these modernization projects usually run way over budget and
>>>> seldom accomplish their original goal.
>>>
>>> Huge IT projects are in general risky.
>>>
>>> Migration projects are no exception.
>>>
>>
>> Which is all the more reason to stay the course and clearly
>> understand "modernization" before you start throwing terms
>> around.  A COBOL IS running on a PDP-11 or TOPS system does
>> not need a new language.  Re-writting it in Java or C# or
>> even Python will get you nothing but a potential for new
>> bugs, inefficiencies and business logic problems.
> 
> Not true.
> 
> It will get you on a supported platform where you can
> easily  get people with the skills.

It will also get you a program that was originally written with
a language designed to do the job replaced by a general purpose
language not designed to do any particular task.

> 
> Yes - a migration come with some risk.
> 
> Main risk mitigation factor is the skills of those
> doing the migration.

Skills can be acquired.  Most of the languages used for these legacy
ISes are much less complicated than modern languages.  Complication
was added by academia to show how brilliant they were.  It brings
nothing to the table as far as getting the job done.

bill