[Info-vax] Long uptime cut short by Hurricane Sandy

[Stephen Hoffman]

On 2013-01-27 05:37:42 +0000, Alan Feldman said:

> My boss's boss maintains an XL spreadsheet with all of our apps listed
> with much information about each app. There are maybe two dozen
> columns of stuff for close to 284 apps. I need only 5 of those columns
> to post in the Application Owner Table (AOT) on our Confluence wiki,
> which runs on a Unix system.
> 
> Now, sometimes he has to make changes: add or remove apps, or change
> owners (each app has "owners"). So every once in a while he updates
> his copy and emails it to me. I FTP it to the VAX. I do some very
> quick edits (using EVE, as the lines are too long for EDT!) to take
> care of a couple of things that would be far too time-consuming to
> code. Then I run it through a DCL command procedure (which calls other
> DCL routines). The DCL routine extracts the 5 columns I need and
> converts them to wiki markup. It also adds headers and a short
> informational note. I then FTP the wiki file back to my PC. I open it
> in Wordpad and copy and paste the wiki markup into the appropriate
> Confluence page.
> 
> That's it! That's ALL I use the VAX for. Yes, I know it's depressing,
> and perhaps hard to believe, but that's it. Oh, I also use the SEARCH
> command to check what owners own a given app when I need to know that.
> But that's it. Certainly not worth buying new hardware for!

This collection[1] of DCL command procedures is probably not likely 
worth keeping the existing hardware and software, much less the effort 
involved with migrating that DCL into emulation and licensing and 
maintaining and using that, for that matter.

Something akin to <http://dmcritchie.mvps.org/excel/xl2html.htm> 
implemented in VBA on Microsoft Windows, or some other Microsoft-based 
technology[2] and probably with Python on Unix.  Java, Perl, Lua and 
other programming languages can also all be used, as can a bash script 
with awk or xml_grep or sed.

Rather than manually processing the file format translation, it's 
usually feasible to configure a Unix mail server using a sieve script 
<http://en.wikipedia.org/wiki/Sieve_(mail_filtering_language)> or 
analogous, and automating the whole conversion-to-posting sequence.

Other approaches include migrating the data into a database, and using 
Excel as a front-end to access that shared data directly, and coding 
another tool (in most any recent scripting language) that retrieves 
that SQL data and dumps it to something suitable for viewing more or 
less directly within your target web server.

Further finessing the requirements — or massively over-designing the 
solution, depending on one's own perspective here — consider migrating 
entirely from Excel spreadsheets[4] as your primary data to Trac 
<http://trac.edgewall.org>, Bugzilla <http://www.bugzilla.org>, Mantis 
<http://www.mantisbt.org> or whatever your organization uses for 
tracking tools and bugs and applications, and move both the application 
inventory as well as the tracking for the errors and updates and 
enhancements into the same system.  There are also various commercial 
inventory and asset-tracking and bug-tracking databases around, and 
your organization may have one of these systems already installed.

If you do head toward emulation, then simh will probably work for you — 
I'm testing some tools with that, for reasons not germane here — and 
there are commercial emuation options.

————
[1] Does anybody have a suggestion for a collective noun 
<http://en.wikipedia.org/wiki/List_of_collective_nouns_for_birds> for 
DCL procedures?  A collection of DCL procedures?  A nest?  Wad?  
Accretion?  Scrum?  Heap?

[2] There are better forums around where you can get Microsoft Windows 
or Unix programming questions answered.

[3] Spreadsheets can be good choices for single users maintaining small 
databases, and exceedingly poor choices when multiple users are all 
making changes to the same data.


-- 
Pure Personal Opinion | HoffmanLabs LLC