[Info-vax] History of DECSET / CMS

[Arne Vajhøj]

On 3/29/2020 9:26 PM, Dave Froble wrote:
> On 3/29/2020 8:38 PM, Andrew Shaw wrote:
>> I have some brand new code I am writing in Python for a utility to run 
>> on our VMS environment and it is completely stand-alone from the rest 
>> of our app code and so that will be going into Git which is the right 
>> answer as far as that utility is concerned.
>>
>> There is also a "desire" to move all our app code from CMS into Git as 
>> well, but we have some challenges around that for not really much gain 
>> so I'm not sure that will get much traction. We *could* definitely do 
>> it, but I'm not sure we gain too much out of it, but it still could 
>> happen.

> Got to say, through the years I've encountered more than a few people 
> who wanted to change things, just because they wanted to.  I'd always 
> ask, what are the advantages, the cost, and the disadvantages.  With due 
> consideration, not once was there ever any valid reason to do so.  Not 
> saying it cannot happen, but for me it never has happened.

 > My own opinion of such suggestions is that some people want their
 > employers to support their fancies, and/or their supposed career
 > enhancement.  Never thought that was what employers were interested in.

It happens that people spend time and money on migrating to
something that was "hot" but quickly turned into a dead end.

But the specific move being discussed here is something that
most IT companies did 20-25 years ago. So it is not "this years
fashion".

And I don't think I have ever heard about anyone regretting
migrating from CMS/RCS/SCCS to SVN/Git/Hg.

> It just seems to me that people spend so much time, wasting time and 
> effort.  In this example, users would have to learn a new product, 
> wasting all the past efforts in learning and using the current product. 
> There is a real cost to that.  Then there is the inevitable mistakes and 
> loses.  It's going to happen.  Guaranteed.  Ever wonder what real and 
> valuable work could be done with the same time and effort?

Training cost is definitely something to consider.

But what way it points depend on the type of company.

If a company have the same employees they had 30 years
ago and do not have any plans on hiring new ones and when
the last employee retires the product is dead, then switching
to a modern VCS is extra cost of retraining the employees.

If a company is hiring new employees and plan to keep
existing for decades after current employees retire, then
switching to a modern VCS actually reduce cost long term,
because the new hires know the modern VCS and would need
extensive training to learn the old one.

And then there is the matter of tooling. For something like
CMS then it only works well as is for simple setups - for
complex setups a company specific scripting layer is usually
added. That is rarely needed for a modern VCS. The time
spent on maintaining that additional layer can also be
used to produce something sellable.

Arne