Reply to both (is RR too hard / easy?) and (is RR marketed to dev'rs only)

[Peter Alcibiades]

Two thoughts.

One, a big gap in Rev as a general purpose environment seems to be getting 
data out in a structured form.  Or maybe this is just personal ignorance in 
how to do it?  If so, corrections would be welcome.

You have lets say a tab separated file with data in it.  This seems to be the 
approved way of storing a few thousand or tens of thousands of records of 
data, and indeed it does work fine for storage.  Now you want to produce 
reports out of it.  The recommended way (which I've done) seems to be one 
field per item of information, lay out the card so that it looks halfway 
decent, and then print card.

However any report is quite capable of having 50 -100 cells, more, in it.  
Now, it may be that there is an easier way, which if so it would be great to 
hear about, but addressing all those cells individually in a script so you go 
through the file and extract and calculate what is wanted is immensely 
tedious, inflexible, lengthy and error prone. Also, your reports may well be 
multiple pages, in which case layout again becomes problematic.  The card 
metaphor does fine when you are dealing with screen fulls of info, but when 
you're dealing with page fulls, less so.  Maybe there is a good way to lay 
out a field with totals and subtotals and percentages in it, and that does 
not involve addressing all the cells individually?  Even then however, in 
Linux, I would have to re-export it to a text file at the moment in order to 
format it properly for printing it.  (Yes, a bug has been filed on this.)

Consequently in this, as in formatting a field for printing, your average 
dinosaur sighs and reaches for awk (animals of more recent period will 
probably reach for Perl), while reflecting that this is for at least this 
purpose an incomplete environment.  Not only does it take a fraction of the 
effort and an even smaller fraction of script length to do it in awk, but the 
thing actually looks like a proper report when you are through by default - 
you don't have to mess around with setting the field parameters and alignment 
of the boxes so that it all displays nicely.  But, you feel, this is the sort 
of thing I should be able to just do natively, without going out to shell in 
different ways potentially in three different operating systems with all the 
complexity that introduces.

Could you do it in Quartam?  Probably, but not in Linux at the moment.  

I don't do this for a living, and restrict my OS environment, so its an 
irritation not a showstopper, and its balanced by other eases of use in other 
areas.  If I did it  for a living, and on multiple platforms, it would make 
me think long and hard before adopting Rev as the general tool of choice.

A second issue (again it may be ignorance of best practice) is the sense of 
fragmentation.  As the apps you take on become more complex, the tendency is 
for the code to spread itself across many objects on many cards.  Now, for 
maintainability, I'd like to track all uses of a given variable name or 
object name in every script.  Its not obvious how to do it.  What I do 
(please, tell me a better way) is patiently copy and paste every script into 
a real editor - Geany or Kate.  In fact, I now have started to write the 
scripts in Geany or Kate, having first drawn up a list of objects and 
variables.   Then at least they are addressable as a whole, and in Kate you 
have the double or triple window into the text, and global search and replace 
works.  Its a bit tedious replacing all the scripts when written, and 
tracking what you've done, but its possible given pencil and paper and check 
lists.  It works, but one keeps feeling, surely an environment which is bound 
to lead to lots of bits of individual scripts should have some built in way 
of dealing with this better?  Maybe it does, and I've idiotically not 
noticed?  

Its probably called GLX2 - but again, not for my chosen OS.

If I were 20 and had learned programming starting with Rev, these two things 
would probably be what would make me 'move on'.