[Info-vax] naming convention in VMS MAIL

[Stephen Hoffman]

On 2018-12-25 07:17:10 +0000, Phillip Helbig (undress to reply said:

> In article <pvquq5$lpg$1 at dont-email.me>, Stephen Hoffman
> <seaohveh at hoffmanlabs.invalid> writes:
> 
>> On 2018-12-23 22:06:41 +0000, Hein RMS van den Heuvel said:
>> 
>>> On Wednesday, December 19, 2018 at 4:25:53 PM UTC-5, Phillip Helbig 
>>> (undress to reply) wrote:
>>>> VMS MAIL has files like MAIL$45870BB2000500B0.MAI;1.  The digits are 
>>>> not random, but are related to (at least) the date and time.  What is 
>>>> the exact format?  Has it changed in the last 25 years?
>>> 
>>> Hmm, I don't know why our friend Steve start to waffle about API's, 
>>> opaqueness, performance and such. The question was highly specific. 
>>> "What is the exact format". There is no suggestion Philips wants to do 
>>> something useful with that. Perhaps just curious.
>> 
>> Phillip has a predilection for asking specific questions that are only 
>> tangentially related to the actual requirements and issues.
> 
> Maybe, but not in this case.

Yes, exactly in this case.

Ponder the following hypothetical question...

> Hi, I'm trying to search for messages in my Mail file, and built-in 
> commands aren't sufficient.  I'm considering using the SEARCH tool on 
> the mail database, but the filename formats aren't documented.  Does 
> anybody have any suggestions?   I've tried MLSEARCH and it works for my 
> needs, but it has a few bugs.  I've searched for the filename format in 
> the doc and in postings here in the newsgroup, and haven't found a 
> definition.  I'm running OpenVMS Alpha $version and $other-version in a 
> cluster, if that matters."

And the following hypothetical reply...

> Ah, sure, Phillip, okay, well you or somebody else could fix the issues 
> MLSEARCH freeware search tool that have been encountered, or you could 
> extract the messages via EXTRACT/ALL of each folder and search that, or 
> you could create your own search tool using the callable MAIL$ API.  
> There are other mail clients around that might be interesting, as 
> Netscape Communicator used to be able to access the local mail 
> database, and there are other clients around.  If you use the 
> brute-force directory SEARCH, you can usually figure out what database 
> entry the message is associated with, if the match is in an external 
> file.  But scripting the mass extraction using the MAIL API or 
> less-desirably using MAIL commands is probably a better approach.

You quite often lead with your proposed solution, and include little or 
no background on the problem you're seeking to solve.   This routinely 
gets you an answer, though it's questionable whether you will get a 
particularly good answer.  Asking specific questions tends to draw 
fewer alternatives and options.  Or none.  As some bozo around here has 
commented on occasion, "ask a terse question, get a terse answer."

> 
>> In this case, the only typical reason to be asking for the format of 
>> the opaque filename is the creation of a tool to rebuild the mail 
>> database, and tools for that already exist.
> 
> Then my reason wasn't typical.

Your entire use case is not typical.

I did what you're doing here with OpenVMS on the desktop too, and for 
several decades.  Writing tools and fixing tools and working around the 
ever-increasing limitations—search, MLSEARCH, web browsing—comprises an 
ever-increasing amount of that desktop effort.

In this particular example of the limitations, OpenVMS lacks a caching 
fast-search tool, and which would render this whole discussion moot.  
The fundamental issue being that OpenVMS isn't targeting your preferred 
desktop use, and that the investment in OpenVMS has been lacking.  
macOS and Spotlight (cached search tool and app launcher) completely 
blows away what OpenVMS offers here, and the Windows cached-search 
analog to Spotlight also does quite well in comparison with OpenVMS.



-- 
Pure Personal Opinion | HoffmanLabs LLC