[Info-vax] A DCL wish list of sorts...

[Stephen Hoffman]

On 2019-03-18 13:57:35 +0000, pcanagnostopoulos at gmail.com said:

> On Monday, March 18, 2019 at 9:21:50 AM UTC-4, Simon Clubley wrote:
>> Using an escape character is more readable IMHO. That would allow you
>> to write your example as something like this:
>> 
>> $ write sys$output "Print a double quote \" followed by a single quote \'"
> 
> Yes, an escape character would be better. To add one backward 
> compatibly, we would have to add a declaration:
> 
> $ escape_char "\"

VSI are going to run with what they have for the foreseeable future.   
VSI lacks the schedule and staff and demand for a DCL replacement.

Best short-to-intermediate answer is probably the addition of an 
embedded scripting language such as one of the Python variants.

There are enough other problematic details around DCL—not the least of 
which is the lack of UTF-8 support—that will make starting over and 
migrating (or translating) existing DCL into the new environment a far 
better result.

OpenVMS has a long history of these backwards-compatibility settings, 
and that's what led to the 64-bit addressing APIs for instance.  Try 
creating a 64-bit app.  It's definitely not intuitive.  Where it's even 
possible.  Try using a secure password hash.  There are all manner of 
other broken defaults, such as extended parsing.  Works great for 
existing customers, of course.  Not so good for new work at these same 
existing customers, nor for acquiring new apps and new customers.

Easiest answer here is to keep nibbling around the edges, and to 
continue making the existing messes bigger and making achieving the 
appropriate settings more confusing and more arcane, of course.  Or as 
it's also sometimes known, "upward compatibility".

Not a fun trade-off, this stuff.  For VSI, or for VSI ISVs, or for 
end-customers.


-- 
Pure Personal Opinion | HoffmanLabs LLC