[You are now on a webpage filed as: http://www.danbyrnes.com.au/personal/perslastpage.htm - [Next page (the first of this series)- /personal/personal1.htm

Look for this
Home Page
navigation button
as you travel.
This page updated 29 July 2010

www.danbyrnes.com.au

Contact via the convenient (and virus-free): e-mail form

To blog or not to blog?

This is the last page of this personal website. It's mostly for material arising from the latest in personal business, a new poem, a new remark, and so on.

Upon my sister turning 60

(Poem 799 on 15-11-2009 draft3, written as the sun comes up)

Cooking at a friend's party one night

Upon my sister turning 60,
and upon my soul,
I recall how I turned 60.
Luckily, without incident.
It's like finding a great flat plateau
with not a lot new to look at.
The water is fresh, the animals and the breezes are slow,
the grass is green enough for the cattle we inherited,
and neither of us sings hymns anymore
of any particular description.

We shared a childhood, then life
shared us out into busyness that seemed important,
until surprise ebbed, or mercy gasped with disbelief,
or The Veils of Maya lifted,
the climate shifted.
Conclusions drifted,
because the feeling grows
that life is more gifted
than we can ever unwrap.
And then what?

It's all in how you look at it,
read a book at it, find a nook,
eat what you cook
for yourself,
and discover that God is a skyhook.

You don't and anyway can't tell anyone the secrets,
because each secret reshapes itself a little overnight
as it bellyflops into the past to escape you once again,
and you wake and wonder if sleep is practice for death,
or if death is practice for another very long dream.
If the sunrise is better than usual,
that could be quite enough versus the evils of the day.

The earth turns, but it doesn't exactly go anywhere,
nor does it seem to care about that,
it's just involved in turning,
yearning for more turning,
yearning and turning, yearning and turning ...

Maybe we ended up in entirely free space?
I don't know.
Just that it doesn't feel so expensive anymore, and here, there's much less wear and tear.

Advertisement

More nonsense about monarchical Australia

Item 15 November 2009:
I was very amused watching TV news last week, more so for being republican in outlook in Australia. Someone suggested, while discussion on a republic here (or not) comes and goes, that if Australia is a monarchy (it's a symbolic monarchy after all), why not have our own king (or queen)?
Lots of mirth, here. Why bother? We really don't want to install a ruling dynasty, do we? Really, and without a fight, we would choose some particular family for the purpose? Which family? A family to be given some kind of power (or duty of conducting ceremony) that no other family has? Why not have a republic and just elect a prime minister, or president, who has power for a limited, earlier-stated period. Who is later to be respectfully sidelined to a Council of Elders to advise on future developments?
I'm reminded of all sorts of things. The derision in Colonial New South Wales that was lavished on politician William Charles Wentworth when he suggested the establishment of an Australian aristocracy. Insubordinate proletarians promptly termed this "the Bunyip aristocracy" and chuckled themselves to death. So all Wentworth became was a First Bunyip Among Equals! And a very amusing one, neutered by the growing powers of the Australian sense of humour.

I'm reminded of all sorts of things by this nonsensical monarchy remark. Occasionally we find an Australian politician saying something about how "great" Australia is, or could become. More nonsense here. Australia isn't great, it's merely interesting. Very interesting indeed. Any idea of "greatness" of course would have to be drawn from history. Greatness of that kind, great wealth, great power or influence in affairs is not something that Australia is destined to possess. Our continent is far too old, tired, soil-worn-out and arid to have patience with any such human-produced nonsense. Our cities are also too badly-planned to ever be able to house "greatness".

At a poetry reading in Armidale

	Dan Byrnes at an April 2008 Poetzinc reading. Photo by Bob Cummins

The Australian landscape simply doesn't care about bombastic human notions such as "greatness", and nor should Australians care.
But Australia and the life forms it is home to, the histories it presents, are interesting, fascinating, different, often puzzling, and offer unexpected contrasts to anyone familiar with anywhere else in the world. That is the destiny of Australia and Australians. Not to be great. Simply to be interesting, very interesting.

As a matter of world heritage, tradition and history, no one should ever forget. That until recently, so recently that it doesn't matter anymore, no one ever illustrated the story of Noah's Ark with a kangaroo seeking shelter from the coming storm. That's because no one knew that kangaroos existed. Till 1770, Australia and its people existed apart from what the rest of the world thought was "reality". This is an ineradicable part of the meaning of Australia, and it always will be. This is something that cannot be contradicted, and should be appreciated throughout the world -- that life can always offer surprises.
Meaning, the role of Australia is not to be great. It's role is to be surprising, to present the unexpected. And on the world stage, we happen to exercise democracy in Australia with enjoyable surprisingness. Long may this continue!

So if anyone suggests that Australia should produce its own monarch, the republican merely refers to "the flag currently in use", and leaves it at that for the time being.
All round, I think that Australia will become a republic when Australians feel the time is ripe. And not before. It's obvious for now, and it has been all my life, that the time is not yet.

Some notes on using Ubuntu/Linux

I've been using Ubuntu for several years now, and find that (assuming my hard disk is ok) computing work is steady, unfussy, reliable, everything just works. Which was rarely my experience with Microsoft software and operating system from my early days of using Windows 95. In fact, after resisting earlier badgering from a friend to migrate to Linux from Microsoft due to the need to keep using one particular piece of software, I finally decided to make the migration when one day, Windows XP fell over one time too many. WinXP had taken to doing this to me, just falling over every so often. That's it, I muttered, this will never happen ever again! It's Ubuntu for me! Goodbye to Microsoft forever!

And right now, July 2010, I'm using Ubuntu 8.04 on two computers (which just now are not networked, but they have been, and all that worked well too). Both computers receive their usual Ubuntu updates in the usual way. Everything is still working well.

There are though some issues about using the greatest variety of software that Ubuntu makes available. But firstly, for those who don't know, Ubuntu's Synaptic Package Manager, which installs new software for us when we want to start using it, offers us the use of up to 6396 software packages. Of the 6396 or so packages, I have 1368 installed, or about one-sixth. It is easier to mention what I don't do with a computer, in order to let you know where I might be up to as a Linux user.

I don't do anything with video-handling software, and rarely look at say, Youtube. No video editing. Actually, watching video and/or movies/documentaries on a computer doesn't interest me. My fooling with graphics is restricted to using The Gimp and checking anything coming from my digital camera (for which sort of picture-handling, Picasa is also very handy).

I am not into any sort of wireless computing scenario. I have no mobile phone-cum-camera. Not into texting or SMSing or Facebook or cloud computing. Not into using online storage set-ups. Not into computer games. Have no useful knowledge of electronics or computer hardware. I'm not into programming, so I don't need new packages of Python or Perl or any other programming languages unless indicated to be so by an expert for a specific purpose.

I do need any available updates of Open Office for word processing and spreadsheet work. Any variety of packages for work with websites I'll look at avidly, such as Bluefish (a robust code editor which I lately use).

I should be using improvements/add-ons to Firefox more than I do for work on website pages. I do handle a lot of sound files for home entertainment, and I've tinkered with music notation software. Using music notation software is a whole ballgame in itself. I've experimented with it, but tend to fall in a heap with the complexities of software for MIDi etc. Also because some music-writing software requires complicated adjustments to a computer's ability to generate and distribute sounds to a set(s) of speakers. There are some strange things with using specialty music-notating software, which happen with usual computer settings for sound-handling, such that I've given up pursuing such musical curiosity, at least for the time being.

And as Linux users and Linux explorers will know, before Ubuntu/Linux appeared, the main problem with Linux systems for non-programmers was with understanding and learning how to do command-line work to get Linux software packages properly installed and actually working. Linux users who were up-and-running ok might well have advised us how-to about certain topics on forums, but they tended to be excessively nerdy, and so were incomprehensible to the non-programmer. I find that this problem - conversing in overly technical language and terminology - still afflicts Ubuntu users via the Ubuntu forums, as follows.

Family History

	"Me - working hard": A 2010 photo of one of my older cousins, Mike Kelly, a computer programmer of Sydney

Solving a computer-using problem has much in common with problem-solving in science generally, and with engineering-type problems - it helps to be very precise with language, with approach to the problem, with concepts and with defining what a problem actually is, or how it arises or shows itself. Relative precision of language versus technical incomprehensiblity is another story yet again. And I struck some problems with trying to do Optical Character Recognition (OCR) work - text scanning - with an Ubuntu system (V8.04 to be precise). As follows.

In my Microsoft-using days, years ago now, I had some quite acceptable software for OCR work. At one point I was using a well-known sort of scanning unit, and quite good, a Canon CanoScan LIDE20 stand-alone scanner for either pictures and/or text scanning. But in 2010, we still find that still being lazy, Canon still can't/won't provide drivers for Linux systems for their Lide20, so I can't currently use this unit.

Going back some years, once I'd adopted Linux systems, I bought a well-known all-in one printer-scanner, unit, a HP C3180, which is generally a deftly-operating little unit, quite ok for my purposes. But neither do HP provide drivers for Linux for OCR work. Or if they do, I sure can't find them on the Internet. What to do? What to do when due to their laziness and lack of curiosity, allegedly world-class software developers and equipment manufacturers can't get off their technical backsides to do a bit of useful Linux programming, and aren't made to by governments!?

Visit forums discussing the problem anywhere on the Internet that I can find them - go google on "HP drivers C3180 OCR Ubuntu text scannning" and so on. No useful clues arise, you'll find. What I did encounter was a lot of complaint on Linuxy forums about a similar problem, inability to get a Linux and/or Ubuntu system to do OCR work. But we find mention of a package named ImageMagick, which I downloaded (and so far haven't used, not as far as I'm aware, anyway). It needs to be used in conjuction with Tesseract, supposedly a highly-accurate OCR worker. So I ensure I've downloaded Tesseract too. But neither ImageMagick nor Tesseract show up on any revised list of Ubuntu applications. So how do I find out where they actually are on the hard disk? Is it worth bothering to try to find them in root? Do I actually have them on board? ImageMagick can (should be?) be used with Tesseract-OCR, but some triers here have had problems here of continuous versus discontinuous output (whatever that means). Forums give the impression that the scanner, via use of XSane, which Ubuntu provides, will provide a graphics file of the scanned piece of text (a .TIFF file is usually recommended due to compression ratios etc) but it isn't then clear what to do next. We are given command line suggestions such as:
tesseract ocr.tif $1
gedit $1 txt

Which suggests to me that somehow, the desired text will turn up (in ASCII?) in the Ubuntu notepad software, which is gedit. A result I still can't get. Then, some forum-poster has tried to use a .png graphic file for this trick, though not so successfully.

I find one poster has written, "I started using Ubuntu several months ago, and was very disappointed with the lack of a GUI OCR package - being a newbie and a non-code-type of soul." Just like me, and very disappointed with it. What on earth are we supposed to do? In Ubujntu's bindles of software, XSane looks like it ought to be the said desired OCR GUI, but things aren't this simple at all.

Suddenly I remembered a discussion years ago with a young IT student living in my university town, something about handling PDF files in Ubuntu (which can be read with Viewer, an unfussy, and totally-free, Linuxy equivalent of Adobe Reader). He suggested, if we want to copy text from a secured PDF file, to use a package called Evince. Which I recalled, could highlight-then-copy the said text. Not that I've experimented much with doing that. But if I tested that option many months ago, then I should have downloaded Evince at the time, meaning I've still got it, presumably.

This seemed to be the missing realisation. So I re-inspected XSane (image scanning software) for the umpteenth time. XSane gives us an option to scan-copy a page of text, which my printer-scanner will do. Ok. So what else? Using XSane we can scan for Viewer, Save (which I'd mostly been trying to do and not getting anywhere useful), Copy (as in make a photocopy, that works ok). Multipage. Fax. Email. If we get a result (that is, a file), it can be rehandled by-extension, or as .JPEG, .PDF, .PNG, .PNM, Postscript, Text (is that .txt?) and .TIFF.

Hmm, I thought, Evince is a viewer for PDF files, it comes up automatically when I click on a PDF. But what is this, Evince isn't listed in any list of software I can find on my computer, as kept by the computer (except in Synaptic Package Manager). Can Evince, perhaps, if we try to use Xsane for the Viewer option, read a newly-scanned image of some text? This question was inexact, maybe even misleading in itself, but, I made a new .TIFF image of some text, put it to Viewer within XSane, the image resulting from XSane's scan is presented within Viewer, hit the OCR button on Viewer (that never happened before [?]), and success, suddenly in gedit, there was my scanned text, ready for proofreading/correction. How all this finally happened, quickly, I'm not at all sure. If it had ever happened before, if scanned text had become available from earlier experiments, I wasn't aware of it. And for example, was this good result suddenly due to some package I'd not long downloaded? There was now no way to tell the difference.

The OCR results meanwhile are still not as accurate as I'd like, but maybe I can refine the accuracy aspect of the OCR work in due course, I've at least got some traction on the problems. And none of the forums suggested, try things in Viewer in the XSane list of options. Or any such tactic.

By now, I think it would be good if Ubuntu could arrange three things:
(1) Provide an option in System - Preferences - Preferred Applications to refer to software to be routinely applied to OCR jobs;
(2) Put some extra checking options in the Printer Config box to the same effect;
(3) For godssake, put some notice in official Ubuntu websites, and on the forums, about how to approach the set-ups for OCR work done with various versions of Ubuntu, depending on the version of Ubuntu being used and any different software used in different versions, where this might be relevant. Because the forums are worse than useless, and mostly seem to refer to failed attempts to solve the problems by the command-line approach.

And now I must go, to see if I can improve the accuracy of what I can get so far. This has taken a long time. There's a lot to be done. -Ed
Evince on my machine cannot be file searched-for in user's folders, so it must be somewhere in root. It's Evince Document Viewer 2.22.2, using poppler 0.6.04 (cairo) - whatever that is. And whatever it is, the Internet informs that poppler is pretty damned complicated. Like you or I do not want to know.
A friend now looking into OCR iwth a version of Ubuntu reports that he has just looked at a review of one release of Ubuntu which remarks about OCR - "Don't even think of mentioning it". Which rather looks as if Ubuntu really doesn't want to discuss OCR. More on other reactions when they come to hand. - Ed

Next steps. A friend visits, someone familiar with Ubuntu but mystified with Ubuntu's OCR ways and wondering about tesseract-OCR. Thinks he might stick with MS systems and Omnipage. But our conversations indicate that Ubuntu has developed two separate camps for approaches to OCR work. One is command-line driven (especially re tesseract), the other relies on GUIs (graphic user interface). These two approaches seem to depart from each other with increasing speed, the more one finds out about them, but there is no clue on how the software is actually behaving differently or not with the use of either approach. It's all a pest. We find we have to use TIFF format for images to be handled by an OCR engine, bu the output file extension has to be .tif, minus one F. It seems that OCR options for Linux systems are driven by pedantic, uninformative fools! We'd sure like to be getting better quality output of whatever OCR work we can so far manage!

Who and what is www.abbyyusa.com. A pay site? It advertises several options for OCR work but like so many shy websites gives no pricing information for the newly-arriving punter. Still, it does offer opportunity to download (free?) for test OCR sessions.

- Dan Byrnes (otherwise indicated in these pages as -Editor)

[Top of Page]

View these domain stats begun 18 December 2005