[Oberon] N-O user's first impression of A2.

Chris Glur easlab at absamail.co.za
Sun Oct 3 03:52:47 MEST 2010

Obviously a long-time N-O user, will first run the familiar 
'Oberon' application of A2, eg. to mount partitions.
And the repeated 'unknown alias <partn>' error message 
makes one believe that the CD has a faulty Oberon.Text , 
and perhaps it has, if A2 does not use the same Oberon.Text 
as the embedded-Oberon?
Or is it known that the mount facility of N-O doesn't work
under A2?

It's a pity that pasting between A2 <--> Oberon seems not possible.
Human Computer Interfacing is difficult to do, and near impossible
to describe/criticise because it entails a multitude of trivial steps.
It's too much like teaching an infant domestic behaviour and
how to tie his shoe-laces. We are embarrassed to spell out 
apparent trivialities, in excrutiating detail.

The following trace/log of steps taken to do a task, shows
just how brilliant the N-O design is, compared to others;
and lists the problems that N-O does NOT have.

The TASK :
edit and copy the text from Sgt40:p48:System.Log in N-O format
to Sgt80:p1 in FAT format, in order to access it by LinuxETHOberon,
to be able to mail it when I again get access to internet, via LEO.

Firstly, note that the source and destination:
<device>:<partition>:<File> are problematic to remember
or keyin.  N-O allows retaining these <longNames> from
the previous session, when they were worked on.
I.e. an identifier need only be [heuristically] designed once;
and later, only needs to be recognised and 'selected'.

The TASK's algorithm is:
1. mount & access the source-file
2. mount & access the destination-file
3. copy/move part of source-file to destination-file
4. make manual edits.
5. save both updated files.

For 1 and 2,  N-O would have the FileIDs with the 
appropriate <mount> command available from the
previous time, and just need 2 klux!

A2 needs:
remember/find the <device>:<partiton>, possibly with the
aid of -> Files -> Partitions 
-> scroll the partition window to find which device has
Prtn48= Aos-type   .................

Once the partition is mounted, the scrollability of the partition
window, disapears. So in order to mount the 2nd partition, I
have to close the 'panel' and re-open it -- via 3 links.

Whereas N-O allows the user to select a heuristically
valued ID for the partition name; A2 allocates its own name;
so that later you don't easily know/remember if Auto1 is the 
N-Osource or the FATdestination.

Then when you access the files, whereas N-O would just
have them 'listed' from the previous session [I guess A2
can do that too, if you keep logs], A2's <FileFinder> seems
not to have <order by newness, instead of alphabetically>,
which is a must. You don't want to have to remember the
name, and scroll in 400 files, when you could just recognise
the name in the top of the stack of most recently accessed 

And most importantly, replacing the orderly 2 track, well
lined-up layout of N-O, with a simulation of real-life messy
desktop of randonly dropped papers is a disaster for me.

Also the apparently unlimited set of bakups, created by
<store> seems inappropriate. Now I've got to go and delete
them.  Perhaps WinXYZ users don't bother with these
'background details'?

And <textString paste> seems to have some work-aroundable,
difficult-to-describe errors.

Where's <FrameGrow> and without <FrameCopy>, how do
you: read the text with the <link numbers> and extract the
corresponding numbered:URLs which are 6 screens down,
to paste to another file ?

I spent a few minutes analysing WHY the A2 PartitionPanel
was so annoying. It's because it takes screen space, whereas
the N-O method just 'fits the required command in with the
rest of the text on the screen'.

Look at this disasterous output from OFSTools.Watch 
| Auto0: AosFS on IDE0#7
| 	38MB of 132MB free
| Auto2: AosFS on IDE2#48
| 	626MB of 760MB free
| Auto3: FAT 32 on <no name> (read-only)
| 	759MB of 1004MB free
===> I can't see which device:partition Auto3 mounts !

So, every system that I've tried is inferior to N-O.

A2's PET/IDE looks interesting, but I first need a rest. 


== Chris Glur.

More information about the Oberon mailing list