January 2009 - Oberon - mailman3.inf.ethz.ch

Does Oberon/Aos/BB read mime-email ?
by Chris Glur 20 Jan '09

20 Jan '09

Have we got any got any source code to clean such garbage: "enneth=0D=0APollack on Iran, talked about this.=5B3=5D The problem is t=" ??! Apparently this fad/curse is called "MIME-Version: 1.0". Since neither Mail.Mod nor MIME.Mod refere to "=0D", "=0A" nor "=20" is seem that they don't handle this curse. Has AOS/BB been forced to follow the fads more closely ? TIA. == Chris Glur. dumb-web-promotion is killing Usenet and [non-html]-email.

1 0

RE: [Oberon] Texts.Reader question
by Treutwein Bernhard 16 Jan '09

16 Jan '09

Hi Sascha, > What happens if I open a Reader on a Text, read some data, then > modify the Text (e.g. by deleting a piece of text containing the > Reader's current position), and then read on without reopening the > Reader? Will the Reader operate on the modified version of the Text, > or on the version as it was at the time of OpenReader call, or is as far as I know, you get back the modified version, but don't take my knowledge as granted ... > In what document can I > find this aspect described? I tried reading the "Project Oberon" > book, but I didn't find any mentions of this situation. you should find it in Martin Reiser's Oberon System. regards -- Bernhard Treutwein Ludwig-Maximilians-Universität Ref. IIIA3 Anwendungsentwicklung Martiusstr. 4 80802 München Tel. 089 2180-2774 Fax. 089 2180-992774 Mobil. 0152-01549335 e-mail: bernhard treutwein (at) verwaltung uni-muenchen de > -----Original Message----- > From: Alexander Iljin [mailto:ajsoft@yandex.ru] > Sent: Friday, January 16, 2009 8:18 AM > To: oberon(a)lists.inf.ethz.ch > Subject: [Oberon] Texts.Reader question > > > Hello! > > I've got a technical question about the Oberon Texts module. > What happens if I open a Reader on a Text, read some data, then > modify the Text (e.g. by deleting a piece of text containing the > Reader's current position), and then read on without reopening the > Reader? Will the Reader operate on the modified version of the Text, > or on the version as it was at the time of OpenReader call, or is > the behavior undefined in this situation? In what document can I > find this aspect described? I tried reading the "Project Oberon" > book, but I didn't find any mentions of this situation. Is it > prohibited to change the Text being read, or will the Text notify > its Readers somehow of the change? > > ---=====--- > Alexander > > >

1 0

Texts.Reader question
by Alexander Iljin 16 Jan '09

16 Jan '09

Hello! I've got a technical question about the Oberon Texts module. What happens if I open a Reader on a Text, read some data, then modify the Text (e.g. by deleting a piece of text containing the Reader's current position), and then read on without reopening the Reader? Will the Reader operate on the modified version of the Text, or on the version as it was at the time of OpenReader call, or is the behavior undefined in this situation? In what document can I find this aspect described? I tried reading the "Project Oberon" book, but I didn't find any mentions of this situation. Is it prohibited to change the Text being read, or will the Text notify its Readers somehow of the change? ---=====--- Alexander

1 0

Re: Power-down restoration
by Chris Glur 05 Jan '09

05 Jan '09

John Drake wrote:- > System 3 in general supports a persistent desktop idea. > I assume N.O. does the same. Assuming each document > you had opened has been saved it will bring you back to > where you were. But that's for the S3 windowed environment > only. Yes !! I'd forgotten about that. That seems to be perfect. I don't use desktop mode. IMO it's part of the feel/experience fad, vs. actually productive facility -- more below. Can you see any reason why adapting that to the 2-track model would be problematic? AFAICS the desktop's [Store] is the 'soft lander' which allows you to power-down without loss of vital configuration info, to just continue your tasks at next power-up. -- snip -- > It certainly sounds doable. But I suspect you'd have to rebuild the > kernel. Count me out of any kernel meddling ! I want to save energy & my effort. VHL/Naive programming, like EditKeys.Tool gives big results for little effort. Imagine, you can easily build a multiLanguage tutor/template/how2getStarted for bash, forth, python ...etc., where to get the python-WhileDo template, with embedded tips according to your wish & colours, you just enter: pw<insert>. Any textual-leaf of any classification of texts is immediately available via the minimal mnemonic keystroke/s. BTW with LEO & VNC & my evolving spell.Tool , I can spell check this S3 email text ;-) The linux & S3 way of just adding extra modules is good ? == Chris Glur. <I'm saving this rant here, till I find how to tell MIT. MIT, which I had considered as the top tech-institute [ETHZ is 2nd ;-] has a java base 'constraint analyser' GLP project. alloy(a)mit.edu I'm a softy for any program productivity tools. Also I've been watching the OLPC project closely, which I 'associate' with MIT too. And again my original crit, apparently confirmed by their change, is that they didn't put the power-consumption at the top of the agenda. I noted the poorly structured tutor and the resulting dumb kiddie-like queries on their blog-thing. IMO an effective communication mechanism between collaborators is essential for any open-project. Eg your simple prompt towards the existing S3-desktop-persistence-facility is very valuable. By contrast, for MIT, after paying for 6 dial-ups to change/find my password, and wade thru multiple art-student-created web pages, I still can't post my techno-query. Ie. the substance is lost in the layers of arty style noise. This whole problem, on a global scale is manifested in the fact that only the technical [and disciplined] News groups still thrive. Where multi-step discussions accumulate knowledge. The rest [apparently the majority] has gone off on to the web-based 'feel/ experience' where apparently they use 'social-web' to post pictures of their cat. I read that the US education system has degenerated ?>

1 0

Re: Power-down restoration
by Chris Glur 05 Jan '09

05 Jan '09

BTW my motive for origniation this topic was NOT how to restore after an unintended power-down; but how to be able to restore to not lose the work-continuity [especially to mental model of where all the pieces are located] after an intentional 'soft-landing'. It's about energy saving. > > I've often wondered how difficult it would be to > > restore the arrangement of the tree of User & System > > viewers under Native Oberon. John Drake wrote: > I'm always up for learning something new. > --snip-- > From an Oberon perspective it would be interesting to see this > feature implemented in BlueBottle and if any of the developers > were interested the link I gave would be a good place to start for > ideas. AFAIK for N-O all that needs to be saved is the display details. Other globals like search & replace strings don't need to be saved. I guess the set of loaded modules, will just be individually reloaded on demand ? My idea of how such a project would evolve [by successive refinement, is:--- Find the root of the User & System display tree, either from a given global, or by linking up towards the root from a visible frame. Trace [output] the Frame labels while linking. A possible spin-off might be the easy ability to: 1. list the total set of frames, 2. select one to be 'moved to the "top" ' 3. rearrange the links accordingly. Coming down from the root, would include all the embedded gadgets too. Then the mapping of the DisplayFrame contents to the heap would need to be saved too. Perhaps if you just saved the heap and the stack-frame-sequence, that would remove the need to save pieces. This is a non-trivial task !? == Chris Glur.

2 1

Big productivity boost with LEO.
by Chris Glur 05 Jan '09

05 Jan '09

Some years back I 'trolled' that ETH insiders were 'hiding the productivity tools under the table'. I got no response. Nor did I get much response about my investigation into EditKeys.Tool . IMO EditKeys.Tool is a massive productivity booster. My analysis of why many linux projects don't progress [for me] is that the parts: tutorials, own notes, Usenet queries ..etc. get dispersed. A strong example is the nice public-domain AI language package 'poplog', where I always needed a half-day retraining to start using, after leaving it for 6 months. It's the trivial clerical tasks that matter, not the hi-tech. With Linux-ETH-Oberon and VNC you can have a *.Tool for each project where all the clerical details are available. Using the principles of 'just filling in the field' & that a convenient programming language doesn't need more than IfThenElse, WhileDo constructs [I've added Case because it fits with little effort], I'm playing [again] with forth. So here's another application where EditKeys.Tool shines. It's trivial to write the macro so that "w<insert>" is replaced by:------------------ BEGIN \ WHILE code1 n \ bool flag WHILE \ DO code2 -(n+1) \ Stmt REPEAT \ WEND balance Stak cr) -(n+1) -------------- And the reserved words are BOLD-red of course. And you adapt/evolve it as you use it. So instead of grovelling, constructing with chars, you build with BIG-blocks. -------------- With VNC enabled my gforth.Tool has the following mess as its 'contol panel':- <><><><><><><><><><><><><><><><><> EditKeys.Reset EditKeys.Read gforthMacroKeys ~ run gforth from VNC by gf6a gf-tut= chroot /mnt/p6 info gforth /usr/local/info/gfWrdIndx ET.OpenAscii ^ Pert /usr/local/info/gfConceptIndx /usr/local/info/clfNews gf6a Menu.fs -e bye == fails to ld Menu.fs after starting. include Menu.fs <-- ok to ldFile after started /usr/local/share/gforth/0.6.2/Menu.fs chroot /mnt/p6 info gforth <><><><><><><><><><><><><><><><><> which takes on a life/meaning to the originator as he colours each part/word in a mnemonic way. And the various notes and tests are further down the *.Tool , separated by "<><><><><><><><><><>", with the contents of the 'book' written on top. So you can skip to eg. topic : "How to chunk stak-assertions between <brakets> ?" by select & [Search]. So each linux project is centraslised under LEO with its own *.Tool. And you can have seperate keyMacro sets for each project. And you can even have the KeyMacro manage a heirarchy/menu of projects/keyMacros. == Chris Glur.

1 0

Power-down restoration
by Chris Glur 04 Jan '09

04 Jan '09

> > Altho' I don't pay for electricity, society pays for me. > > Since 'laying out the various bits & pieces to familiar locations' is a > > major task, in increasing productivety in eg. Linux, a killer app. which > > is long over due, would allow a power-down and reboot to continue > > from the same 'layout'. I suspect it may already exist ? > John Drake wrote: > It's been done before. See: > > http://eros-os.org/eros.html > http://eros-os.org/essays/Persistence.html > http://eros-os.org/essays/reliability/paper.html > Yes but we don't want to learn another OS. I thought I read of an app. for linux ? I've often wondered how difficult it would be to restore the arrangement of the tree of User & Sytem viewers under Native Oberon. == Chris Glur.

2 1

Re. Power-down restoration
by Chris Glur 03 Jan '09

03 Jan '09

> > Altho' I don't pay for electricity, society pays for me. > > Since 'laying out the various bits & pieces to familiar locations' is a > > major task, in increasing productivety in eg. Linux, a killer app. which > > is long over due, would allow a power-down and reboot to continue > > from the same 'layout'. I suspect it may already exist ? > W B Hacker wrote:- > For more than 20 years.. > > The 'killer app' if you will is a side-effect of virtualization, and is > effectively cross-platform. > --snip -- > Finally, these containers often open to an active state faster than > their ordinary 'bare-metal' host could boot and open the apps to the > same level of usefulness. > > Thereafter the 'killer app' is whatever rocks your boat. It's a killer-app, in the sense that a warm coat is, during unanticipated cold weather. Do any of the current WinTel-wallmart boxes do this ? I was hoping that someone would point me to Linux facilities, since I had a vague feeling that it already existed. Although since it would have to work rignt down at the kernel level, it's not just a trivial add-on. > > Will it be economic to have a home-server, when boxes can be shutdown > > and rebooted to the 'same state' by saving the relevevant RAM conditions > > to disk ? > Joerg wrote:- > What you describe here is called persistance and John Drake pointed > out possible solutions to it. > > However, I think SERVERS are a little "special": They don't need > (complicated) persistence. I mean, they sit there on the internet > and wait for client requests. After answering the request they > sit there and wait again. > > When a server boots up, the configuration files will start all > processes that wait for requests. The only thing I had to do > on my server was to program the computer's power controller in > a special way: If the power is cut by accident and will be back > again after a while, I had to make sure that the server is re-booted. > As you know, computers won't boot if you just plug in the power > cable ;-) I'm asking a socio-economic question & Ya'all giving only technical p.o.v. What is a typical duty-cycle of a home-server ? What if each year there's an extra 20 million Chinese home-servers running 7/24 ? == Chris Glur. and related:-- Newsgroups: alt.comp.software.newsreaders,news.software.readers,comp.os.linux.networking Subject: blogs kill groups; cars kill trains; unsustainable ? Except for the most technical, newsgroups have all migrated to blogs. So my ISP newsServer is often disabled/sluggish. When I try blogs, I see that 'the whole book' is transfered back&forth each time an addition is made. And of course there's no concept of threading, which gave further specialisation and opportunities for bandwidth economising. So then I read about RSS, which 'activates on new/additional stuff'. But does this allow Txing of only the new material, or is the whole book still Txed ? What do they use in China/India ? I can't believe they'd use such 'convenient'/wastefull methods as we obsese slobs do? == TIA.

3 3

Re (2): [Oberon] Re: Web servers
by peasthope＠shaw.ca 03 Jan '09

03 Jan '09

Bill, wh> ... a virtualization-engine ... Qemu is an example? ... Peter E. -- http://members.shaw.ca/peasthope/ http://carnot.yi.org/ = http://carnot.pathology.ubc.ca/

2 1

Re: Web servers
by Chris Glur 01 Jan '09

01 Jan '09

Joerg wrote:- > Your hosting provider (not the ISP) is www.verio.com. If you have your > own server at home you won't need that any more. Will it be economic to have a home-server, when boxes can be shutdown and rebooted to the 'same state' by saving the relevevant RAM conditions to disk ? Altho' I don't pay for electricity, society pays for me. Since 'laying out the various bits & pieces to familiar locations' is a major task, in increasing productivety in eg. Linux, a killer app. which is long over due, would allow a power-down and reboot to continue from the same 'layout'. I suspect it may already exist ? == Chris Glur. PS. 2 more trivial tasks that LEO + vnc allows you to do, with the the look & feel of NO: * spell check, * someone complained that my NO email date fields were not RFC compliant [NO's Sytem.Date 's output is not RFC compliant] apparently linux's `date` output is. Previously I'd have to paste it into a linux file and read that with LEO, which is too much bother. Now with LEO + vnc you can see the field to be copyPasted all on the same screen.

4 3