[Oberon] cooperative bug-fixing and refining & underscore

[easlab at absamail.co.za]

--- easlab at absamail.co.za wrote:

> > Well I just see a list of directories which I can't
> > enter with
> > the 3 (successivley refined) methods which I used to
> > approach it:
> >  WebVia email, Desktops.OpenDoc ^ , Linux: lynx !!
> > What happened to "any browser" ?
> > Any thing only accessible by (non standard)
> > Bluebottle is not for me.

John Drake wrote:
> I think you misunderstand what is going on here.

Correct. I still think that http is (should be) for hyper-text.
And I'm reminded of this when n-o has problems fetching
http*.tar.gz  or  http*.ps ...

> This is a WebDAV repository and hence it's accessible
> by any standard WebDAV client.  BlueBottle has such
> a client.  It should be possible to port the WebDAV
> client to Native Oberon, but that isn't complete yet.
> As Edgar mentioned in an earlier mail WebDAV requires
> XML.  I'm the only one working on porting XML to
> Native Oberon (to my knowledge) and I haven't looked
> at that for some time now. 
> 
> Also my personal view is that N.O. enthusiasts should
> at least look at BlueBottle since all N.O. code I've
> seen so far works under BlueBottle fine.
> 
> At any rate an online co-operatively maintained
> distribution of Oberon will require a versioning
> system of some kind.  CVS is still the most popular
> version control system (it powers SourceForge) but
> there currently is no Oberon based CVS client and
> I don't see one on the horizon.  WebDAV is a strong
> upstart and there is an Oberon WebDAV client,
> although currently it only works with BlueBottle
> and WinAOS.

You make a convincing argument !

Edgar wrote:-
> > > So why do you hesitate to give Bluebottle a try ?
> >  >No time ? 
> > > No Pentium processor ?
> > 
> > Both: observations suggest that Bluebottle will fail
> > the cost-benefit
> > test.  Have I told about "the boys that are always
> > polishing their
> > vehichle, but never make a proper journey" ?   
> > Nativeoberon is a work-horse for me, not a toy.

> I suppose this makes a big difference as to what
> part of the world we are talking about.  Here in
> the states we literally "throw away" Pentium 100s.
> And Bluebottle is out of the box capable of doing
> "real work" out of the box, not just "toy" stuff.
> For me the ability to do XML is huge.  I've seen
> very little cost and very much benifit to 
> BlueBottle.  When WinAOS came out my only 
> motivation for porting XML to Native Oberon
> (so I could use it with PlugIn Oberon) went
> away.  I'll look at it again when I have time.
> If there are N.O. users interested enough
> in XML and/or WebDAV to help with development
> I'll look at it again sooner.
 
No, no, 'real life' is different.
Installing a free-gift Pentium would cost 2 days.
Due to a change I have no access to a printer which I previously used to
print 10 pages a month. I though a good brisk walk to the 'print shop'
would be welcomed exercise.  Three days later it's still not working:
*  the printer box is M$ and detects/blocks all my diskettes which
    come off my M$ box with russian-flag virus.  Another 'project'.
*  the margins and page numbering is not good.
*  etc ...   talking is differerent from doing.
==========================
John, you wrote:

> > ...   There is an easier way to change the
> > default font that doesn't require recompiling 
> > anything.  In Native Oberon system variables are
> > kept in the file "Oberon.Text".  Just change the
> > line:
> > 
> > "DefaultFont"=	Syntax10.Scn.Fnt
> > 
> > to
> > 
> > "DefaultFont"=    Oberon10u.Scn.Fnt
> > 
I'm guessing that you have the 'Win' version ?
It's slightly different on non-Win versions.
> 
> Vasile wrote:
>   I guess, for this to work smoothly there must be also modified bold
> and italic versions of default Oberon font. And with correct names.
> This can be done.
> 
>   Besides I was assuming that there is a mapping (from ASCII to 
> typographic signs)
> 
>   - (ASCII minus)       ->  m-dash
>   _ (ASCII underscore)  ->  n-dash
> 
> and that just modifiying the Oberon font will affect the way hyphenation
> is done. Well, looks. Now, after a bit of of trying it seems that both
> Script.Mod and Hyphen.Mod use - (ASCII minus) for hyphenation. So, does
> anyone know which is the use of "Oberon underscore"?
> 
This BAD-HYPHEN is THE repeated example of poor coordination in the 
n-o community: 
* 5 - 10 years ago when I started, I raised the issue,
* every year somebody raises the issue,
* next year somone will again raise the issue,
* it is listed with a patch in some of the documentation,
* Oberon10.Scn.Fnt which is very pleasant is perhaps a special 'classic',
      and when you make your own underscore, you don't know eg. the 
      correct length. A perfect classic is not perfect if one pixel is wrong.

I've spent quiet some time extending Oberon10.Scn.Fnt to use the
ORD(Char) > 127 for the french, german and swedish chars.
Sometimes when I read News etc off the inet, I still find some chars
which need to be done, eg. the 'copyright' sign and the "i" with 2 dots.

If we had a proper collaborative repository, I/we could submit the
font(s) for further updating.

Working on ver-2001 (and also I believe on V2.3.6) there are problems as 
the set of 256 chars become allocated bit maps.  As if some structure is
overflowing.    This was a problem when I started using the extra 128
upper chars for Cyrillc - allocating bit-mapt to their 'official' values.  
If no one else fixes it, I may look at the source code one day.

My impression of the 'mapping' is as follows:

* each char of a text is an octet,
+ apparently the 'header' of the text (file), maps stretches of chars to
   various *.Scn.Fnt (and colour ?),
* the various *.Scn.Fnt files map each of their 256 entries to a bit-map.
* using FontEditor.Panel , you can fetch any *.Scn.Fnt  set; edit any
     of the 256 entries ; Store the new/edited bit-map back and when
     the *.Scn.Fnt (which behaves like a 'library' in that the stored update
      is not immediately available) is re-loaded, you see the new char(s).
* FontEditor.Panel has facilities to show what the edited char would 
     look like when the new/edited *.Scn.Fnt is re-loaded 

The correct view of (especially collaborative) software, is that it is
never correct/complete, and it does not belong to an individual.
Various individuals can add-value over time.

Can WebDav and/or CVS acheive this ?
I'm sceptical of technological solutions.
Wirthism is based on a few key, universally valid ideas  - not technology.

== Chris Glur.