[Info-vax] Hiring maintainers for legacy VMS systems

[Tom Adams]

On Friday, September 12, 2014 7:32:41 AM UTC-4, Bill Gunshannon wrote:
> In article <018b2bdd-c4ef-4aec-9d5a-b784140a25e5 at googlegroups.com>,
> 
> 	mcleanjoh at gmail.com writes:
> 
> > On Thursday, September 11, 2014 10:58:17 PM UTC+10, Tom Adams wrote:
> 
> >> I plan to retire in about 2 years.
> 
> >> 
> 
> >> Is it going to be very hard to find a programmer/analyst to maintain a few 100,000 lines of custom code, mostly Fortran, some C, some Datatrieve, some VMS command procedures?
> 
> > 
> 
> > Is that all? Sounds like part time work to me, assuming of course that it's all well-documented :-)
> 
> > 
> 
> 
> 
> Documented?  What's that?
> 
> 

I think it's pretty well documented for a custom one-off system.

It's a table-driven design.

I'd say the main difficulty is that many detailed features have been added over the years, and I think it could be hard to figure all that out.

In a table-driven system, maintenance often amounts to adding new fields to records or giving existing fields a broader meaning, and writing code to implement the fields/meanings.  This can get pretty complex and obscure. Led me see that one can create spaghetti-bowl tables, where every field is documented but it's still relatively hard to figure out the detailed behavior implied.

There's a concept of refactoring, rewritting/redesigning code to make it more systematic after changes have accrued that were no anticipated by the original design, I have never done much of that.