Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
Linux Business

Journal Journal: tale of 2 technologists

goon in response to #10094868

'... wouldn't you agree Linus got a measly sum compared to Steve Jobs ...'

the salesman
different generation, different locale. Jobs was around at the birth of the PC revolution. He's had plenty of time to create products (hardware + software), make mistakes and sell to a large domestic then international market. I dont think Jobs has ever given away code. Jobs has a knack (and the appropriate reality distortion field ) to foster an ideas environment, root out the better ones (for good or bad: read Insanly Great and think Andy Herzfeldt (And another thing)), take a punt and back the idea to the hilt.

For that Apple, Next, Pixar have delivered big bucks.

the engineer
Compared to Jobs and Apple, Linus and Linux are babies. Linux is a product of it's time. Just like in Victorian England where amateur gentleman had the time (and money) to ponder, think, question and execute their way into the industrial revolution, Linus tucked away in his bedroom with a donated '386 and copy of GCC gave heart and life to the GNU suite of tools in the form of the Linux kernal.

This is one big block in the Information revolution that is now occuring. And while Linus maybe currently *worth less* than Jobs the potential for Linux to generate new wealth is staggering.

In Killer App, Downes and Mui argue that moore + metcalf = law of disruption + coase . Linux and the birth of the Internet has in a way directly influenced this. Anyone who can exploit these effects and sell products stand to make $$$.

Linux is a product. How Linus utilises his time, programming and creating or selling: Its up to him.

'... Beyond that, make $$$ by selling some commercial software that people are not willing to write for their own enjoyment or use. ...'

think diesel not ford
or to put in a different light ... produce software that others have yet to think of or cannot do for themselves. Not everyone want's to sit behind a keyboard and have to understand computers. But to think you can make a living the old way, ignoring mr more and metcalf and hope that distruption and coase go away is shere lunacy.

I like to think of Linux as revolutionary as the Diesel engine (which by the way was not patented and possibly led to the early death of Rudolf Diesel) and Jobs as a Ford like figure. Though the analogy would probably be better with Gates.

To me Linus has unleashed the software equivalent of the Diesel engine. I can leave it to your imagination what legacy Gates and Jobs will leave.


Journal Journal: IronPython 0.6 released as CPL 6

Update: how alpha is IronPython?

reading from the readme with the download you can see it's pretty bare.

  • The only two Python programs that I know for sure run on IronPython are pystone.py and the parrotbench benchmark/test-suite. Of course, IronPython is the only Python implementation other than CPython-2.3 that can successfully run the parrotbench benchmark/test-suite, so this shows that the implementation can handle the majority and trickiest parts of Python's core semantics successfully.

and ....

  • Most of the standard Python library is not currently implemented, so it is unlikely that many existing Python scripts will run successfully under this early alpha release of IronPython-0.6.

and ...

  • it's an implementation of the Python language, you should be able to use all of the language constructs that you're familiar with. By default, IronPython doesn't ship with any of the standard Python library, so if you wish to use these libraries you'll have to provide them yourself.

Ed Dumbill whacked up a comment on IP. Come on Ed. If your going to show a bit of code what about showing the steps to reproduce it? Seems Miguel could run it out of the box. Must be my crappy path setup.

mine for what it's worth is something along the lines of ....

  • d:\IronPython-0.6\bin\d:\mono\mono-1.0\bin\mono IronPythonConsole.exe hello.py

but my python paths are out so I cant load pyGtk. So I'll have to fiddle around with path environment variables. Shouldn't have this problem with Linux. I've got installed Python 2.3.n, Gtk, pyGtk, Mono-1.0, IronPython-0.6 on a win2k box. Looks like I'll have to reproduce this on the linux box.

IronPython 0.6 released as CPL
Hackers coding only in Python. Gimme a break. What we're largely seeing is *script* hackers coding in Python. cgi-bin. shell crap. webbots. It's where Python shines. But there's an awfully big collection of code projects that need to straddle the speed of C++ and the dynamicism (to some degree) of higher-level languages. And there's a lot of hacking opportunity there. Java does that region very well, thank you.

might be time to try IronPython (~1.7 times faster than standard Py2.3). Here's some performance benchmarks comparing cpython to ironpython for those interested in the numbers. interestingly, IronPythons creator Jim Hugunin designed Jython.

Thanks for prompting me to look as Jim just completed a talk at OSCON, 28 ~ IronPython: A fast Python implementation for .NET and Mono and has released the source.

btw it's released under CPL


Journal Journal: rewriting, 'making the news'

...It's not enough for journalists to be mere messengers without understanding the hidden agendas of the messages and the myths that surround it ... john pilger

Dan Gilmour is writing a new book, "Making the News" [Oreilly yet to be published] and asking for RFC. Added some /. comments.

Subject: Ideas for Making the News ~ Ch 2
From: goon
Reply-To: goon
To: j3@gillmor.com
Date: Fri, 21 May 2004 11:34:15 +1000
X-Evolution-Format: text/plain
Hi Dan, I've been reading through the preview of "making the news" and have a few comments..

re: Chapter 2: The Technology that Makes Making the News Possible

I've been on slashdot for a while (uid 2774) and there's a few thingsthat I can observe wrt /. and how *news* can be portrayed. Just like mainstream media, alternative media is still subject to problems. With slashdot problems occur in

*moderation - moderation judgement and bias (caveat emptor)
High moderation gets eyeballs. Moderators can make an arbitrary decision (not necessarily based on how technically/factually accurate a post is) on what they think is a good answer.

So answers to Microsoft questions/posts tend to garner negative moderation for valid responses.

Moderators also gain moderator status due to *experience* not knowledge so the sig/noise ratio can be high depending on what topic a question is asked.

(see www.perlmonks.org for a better example of factual moderation. Talk to merlin aka randall schwartz. read afterslash.org/ to see what I mean about ~ signal to noise ratio over time.)

*story volume - more is (not) better
story volume dilutes the *talent pool*. commenting on stories. I notice that in areas thin in people who actually understand the topic result ina lot of meta discussion on the topic and not facts. I miss reading posts from *experts* in field.

*immediacy - (lack of systematic) follow up and sense of history
/. follows up on stories but nothing like newspaper journalists look for stories. /. tends to cover what others (mainstream media) have covered,but with feedback.

*editors - comment on the story (but don't have insight)
/. editors are pretty green. they get away with murder (sometimes) matching a quick headline and comment to a story and apply 180 to what the linked story actually says. /. ed's are seasoned in some ways more than any journo trained editor for the *tech* insight, implications and technical context.

*story selection - how stories are chosen (far from transparent, can be manipulated)
what is not reported can sometimes be as interesting as what is reported. /. gets the majority of stories from existing media, books or questions from readers.

Stories not reported in mainstream media have less chance being reported on /. unless asked by a lone reader.

*summary (why)
/. is in some way a electronic re-incarnation of the *letters page* in traditional print media. Rowdy, opinionated, funny and sometimes wrong /. is infinitely a more interesting than mainstream press. /. has grown through the cracks of mainstream media expressing it's multicast voice in an increasingly unicast press.

regs PR


Journal Journal: new format journal - longer articles of interest - p1

Upcoming articles of interest
  • Rather than just add links, topics and code without focus. This year I've decided to delve deeper into a particular areas of interest. I tend to post most stuff at slashdot sometimes at my use.perl.com site for perl stuff and occasionally my home page. This year I want to do much the same but filter the topics intelligently depending on the audience.

    Of course I could add an rss link and you roll your own ... one step at a time ... here's some monthly upcoming topics I'm currently writing up ... in no particular order

      • powershift - and how it effects software developers. can it be exploited?
        chefs and development - how cooking shows shed insight into software development.
        desktop, server integration - python on the desktop, perl at the server.
        networks and connections - slashdot friends, foe and foaf.
        stubby - automatic test unit generation.
        slither - simple email and sending complex markup. how POD/YAML saves the day.
        hack the block 1. - why I like (and my kiddies) to use blocks, not just toys. some surprising results.
        hack the block 2. - code visualisation with LEGO
        compiling and aggregating data - what can we learn from map making when compiling data (rss, xml, etc)?


    • building components in high level language python (getting sick of writing business logic in low level code, object pascal, c/cpp).

      • the delphi/python for delphi combination allows you to code business logic in python and the GUI in Delphi/cpp builder.

        andy bulka has a nice explanation explaining the why's and how patterns fit in.

        perl fits into this picture because it strategically gloves apache, CPAN is mature, the language is mature and evolving in a stable manner.

        python integrates into g's better. something that perl sucks at (being embedded). python also runs on windows. something perl via active state tries but has yet to pull off. (see past posts on using ActiveState and CPAN).

        python on the desktop, perl at the server.


    • installed and using sqlite instead of mysql
      • replacing ms access as a generic data bucket.
        not have to worry about running a heavier than necessary server on a client machine
        but now having to think about sqlite restrictions.

    • installed firefox 0.8
      • and thinking about the first firefox in thunderbird, episode 1 ... when I should be thinking of (clints russian flying epic!)
        using firefox->tools->DOM inspector to tackle regexes.
        DOM inspector does that parsing so you can see what tags you need to extract
        why aren't regexes used in mozilla firefox->bookmark manager->manage manager to search for saved links?
        unstable, had a hang with a download and takes longer to save complete pages
        quick mozilla hack - type in, about:mozilla in nav toolbar


The Internet

Journal Journal: 10 years on the Internet 2

10 years on the Internet
    • At the time I was working with a bunch of psychopaths - running around in boiler suits, heavy boots and lugging 50Kg loads in dangerous environments. Taking orders off idiots, having to ask permission to take a drink - but not doing what you are probably thinking. Certainly not enjoying the experience. I had to get out. I was hungry for something and somewhere else to work :)

    Not long ago I noted the 10 year anniversary of the first web page. Well this year marks my 10th year on the Internet. 10 years ago this month, January or February some time, I forget.

    • "... When is not import. Where is ..."

    I had just enrolled to do some some post grad studies. It was at a place of learning that I stumbled into the Internet.

Early Internet - what information can I find?

  • I had to get out of the current job I was in. Doing some extra studies was a good way learn more about computing and software development. The first thing I noticed about the computer centre was the crapy Windows 3.1 machines. Remember this was before the rah rah rah of Windows '95 understanding networking, browsers and IP stack (Trumpet filled this gap). Microsoft had yet to realise the importance of the Internet. Some say they still don't. It took me about one afternoon.

    • "... The biggest change was networking ... We had unlimited access to the web through our student accounts ..."

    Things had changed a bit from my undergraduate years. Gone where the mainframes, punch cards, deck writers. It was light years away from doing mainframe (similiar to this one) , batch card (info for purposes of understanding how batch works only) processing in Fortran, editing mainframe text files in Ed and writing desktop GIS graphical and console applications in pascal during undergrad.

    The biggest change was networking. The computer centre consisted of banks of networked, generic PC's via Banyan Vines to the *nix mainframes hidden deep in the bowels of applied science building. I went there once, to get a new password and was acknowledged by the cliched bearded systems admin who emerged only from their glass-bowled data centre, to quickly read out my 16 digit hashed password. We had unlimited access to the web through our student account.

    Getting onto the Internet involved a choice of CLI or graphical applications. Graphical Internet access was available through a client program called Cello an early browser interface. Mosaic had yet to be installed - had it been released? (yeah by about 2 yrs). I remember spending a bit of time working out how to search with a web search engine, and it's less successful predecessor, gopher and landing in a place called doomgate to grab Doom wad files.

First taste of *nix & internet apps

    • "...access through a mainframe Unix shell dialup account gave me my first real taste of *nix ..."

    Home access through a mainframe Unix dialup shell account gave me my first real taste of *nix. Ftp for downloads, Pine for email, Lynx for internet, vi for editing and c for programming applications.

    I think there was about 50K machines on the Internet by this time. The information space was huge and more importantly, growing exponentially.

    It would a be almost a year and a bit before I started at my first computing job at sausage. But by the end of the first day playing around on the Internet and Web I knew something big was going on.


Journal Journal: m$word-pdf, maestro and australia day

doc->pdf cont.

looks like I have found a possible quicker path to get word docs to pdf. (wonder if it can handle complex tables and embedded images?).

matts (axkit) just left a message on my use.perl journal reporting that he recommends using HTMLDOC for converting html to pdf. It also has an Axkit plugin. I checked out the man and faq. GNU licensing, perl bindings and end user support available.


downloading maestro (38.5Mb) to check out mars rover data. reminds me of the 3d pathfinder back in 97. posted detailed comparison of rover ground images of opportunity and spirit.

australia day

25 Jan. all I can say is crack a tube! (Sound of cans opening) and Australia, Australia, Australia, Australia, we love you amen!.

philosophers song - monty pythons flying circus - Episode 22
Immanuel Kant was a real piss-ant who was very rarely stable.
Heideggar, Heideggar was a boozy beggar who could think you under the table.
David Hume could out-consume Wilhelm Freidrich Hegel.
And Whittgenstein was a beery swine who was just as sloshed as Schlegel.
There's nothing Nieizsche couldn't teach 'ya 'bout the raising of the wrist.
Socrates, himself, was permanently pissed.
John Stewart Mill, of his own free will, after half a pint of shanty was particularly ill.
Plato, they say, could stick it away, half a crate of whiskey every day!
Aristotle, Aristotle was a bugger for the bottle,
And Hobbes was fond of his Dram.
And Rene Descartes was a drunken fart:
"I drink, therefore I am."
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.


Journal Journal: complex HTML-->???-->PDF? 4

the problem

  • but the current Word doc (the catalogue) has tables, graphics and was 'built' with Word templates, so I have no idea how ell it would all convert.

the site that got me interested in pdf was Stas Beckmans site, www.stason.org. He gave a talk to the melbourne pm last year. Through the course of his talk on mod_perl 2 he showed the notes from his site in html with pdf downloads of the site.

So I tried to re-create this html->ps->pdf so that I too could have a printable version of a project I'm working on called Ratpile (make a directory that has *stuff* stored in it searchable by stuffing information about it into a relational database - data mining some may call it.) using perl+DBI+TT2. The template I created is a *bare bones* html page sans images. This is the technique Stas is using with his docset.

the point I guess I'm trying to make is I've used text only and not images. I've done a bit of research and this is what I've come up with...

  • graphics are supported in postscript (3?)
  • others better (ybiC) than I, have hacked together html->PS->PDF code and appears to handle images via html2ps but not html tables (Create PostScript and PDF versions of all HTML files in given directory )
  • one approach could be to use Matt Sergeants, PDFLib (load_image method) a oo wrapper around pdflib by www.pdflib.com. but I seem to remember has restrictions for use under OSI (has to be opensource, private use or researcher).
  • or use Alfred Reibenschuhs - Text::PDF::API where I found via an old page PDF-API2-0 which has some image (jpg,png,handleing capabilities
  • logreport has an interesting set of observations about html->PDF generation. Namely problems with html formatting and tables

building html->PDF with images and troublesome html tables

now given what we have found above I would suggest the following (unless anyone has a better idea) of using:

  • extract word document to html
  • extract table data (word document via OLE) or (via html via Html-TableExtract - like latter better.)
  • remove html tables in html documents
  • reinsert data into a simple table using <pre> tags for layout and html tags for bolding, emphasis. Or find some other method by experimentation in html for representing tables (text)
  • PDF-API2 as the PDF renderer. This can all be done in code.

the real problem maybe rendering the tables generated from word. complicated layout in word (re-rendered to html) will have to be modified to the postscript syntax then rendered to PDF. The problem is defined by converting the html tables to pdf.

it is not rocket science to create a bit of code to extract the data from the table, re-create a table using PDF-API (and its child modules).

but is there a shorcut?

of course you could forget all the above and take your chances with Michael Frankl's HTML-HTMLDOC and convert you html files directly to PDF :)


damn I love cpan.


Journal Journal: Iain Truskett RIP 2003-12-29

update: 13JAN2003 Andy Lester (oreillynet) writes ... I never met him, or even spoke to him except through email and IRC, but I feel like he was a friend. He was always willing to help out when possible, and at least discuss an idea when not.

perlmonks today -

Iain Campbell Truskett ("Spoon"/"Koschei") RIP 2003-12-29

According to his girlfriend's blog, Koschei passed away from a cardiac arrest at the age of 24. He was a significant contributor to the CPAN, and frequently interacted with us on the Perl IRC channels. His wiki is still up for perusal. He will be missed.

Rest In Peace, my friend.

-- Randal L. Schwartz, Perl hacker

Iain will be missed. It was only last November that he gave up some time to present a talk on DateTime at the melbourne.au PM.

Iain talked in detail how the development was proceeding on the Datetime modules and how he spent the last 8 months hard work crafting ("hacking" is a poor choice of words. Conceptually DateTime is not for the weak of mind). The talk was well presented, informative and most of all amusing (especially the bit on the evolution of the French Revolutionary Calander)

I will certainly be thinking of him when I'm using any of his modules. His legacy is more than mere code.


Journal Journal: cop this! 3

should rtfm the /. faq pages before I post.

reading rdf stuff at xmlhack

downloading Kings Quest for the young bloke and playing Hogs of War.

remember to update shipshape stuff. especially since I indexed the ratpile and the rest of the hard drive. created perl code to query all htm* and pdf files used template toolkit to output page and used ghostscript, ps2html and docset to generate pdf of printed page much like perl.apache.org

found why the delay in longhorn and possibly the journalling file system - dependency on yukon (pt 7 managing yukon dependencies). yukon is ms-sql server database in development.


Journal Journal: 'LinuxGazette' + (lawsuit ? '.net' : '.com')

update: 07DEC2003 - did not take too long. Check out SSC vs LinuxGazette.net Continued. Look out for ricks comments.

I'm always on the hunt for online books and references. One site I've thumbed through a lot is LinuxGazette. Filled with a broad range of Linux related topics, programming, sys admin etc.

Then doing some meta-moderation I came across this article, SSC Trademark Threats vs LinuxGazette.net, 12-02-03 9:05.

I originally became aware of this tussle after link harvesting www.linuxgazette.com for later use and checking out their author faq to see if it was worth writing an article or two. Now you know something fishy is going on when the shutters go down, then come back up again sans original crew plus heated boards.

But have ssc bit off more than they can chew? I'm a member of the Victorian Linux Users Group, LUV among others and so is Rick Moen. He actively pops up (even though he's on the flip side of the world) on the LUV mailing list answering technical questions etc, berating those who don't do their homework and helping others who do. I've never really found out why? but myself and others are grateful or at least amused. (I wish LUV's email list was online you could read the posts) The Age newspaper did article on him a while back (SSC used this as a publicity puff peice on their site for issue JAN2003 (see linux links).

stop sniffing my wallet!

In a world of creators and distributors, traditionally it's the distributors who make the cash. Go ahead make your money distributing CD's of documentation, printing Journals but don't take what is not yours - user generated documentation try to spin it into your own cashcow. You're not going to get any documentation from moi at your site.

Sick 'em rick

whats a ternary conditional operator, '?:'.

Star Wars Prequels

Journal Journal: mini-itx + r2d2 + openbds = consumer grade router/firewall

I've been looking at *small* os systems for a while. They don't consume as much power are quiet and you can tuck them into small spaces.

Then I saw this article on daemonnews. Most users (non technical, phb's, grandmothers etc.) roll their eyes at motherboards, operating systems or science fiction especially if you can fit them into 1 sentence.

But if you tell them you have a *secure* internet connection over there in the corner that looks good (insert favourite bit of plastic cover) and they accept it as a *consumer grade* (read desirable) product. It's pretty shallow in some ways but aesthetics in consumer goods sell.

Data Storage

Journal Journal: source/object code on a Lactobacillus cultured CD

sent in entry to ask slashdot. was interested in what would happend to source code cultured with Lactobacillus. does the code still execute? how does the CD reader still read the code (but not alter it)? ....

After watching an intruiging ABC, Catalyst story Design by Fungus about, Dr Cameron Jones who accidently (in the great Australian tradition) spilt beer on a music CD and played it ... the result, the Molecular Media Project which aims include ...

...use cells and atoms to perform useful computational tasks at the micron (10-6m) and/or nanoscales (10-9m) of organisation ... principally concerned with exploiting polymer and colloidal nano-agglomeration or biotechnology to control optical nonlinearity for new media applications ... M.M.P.

The applications of the fungal techniqes have been applied to (among other things) Visual design, audio and images.

  1. Images of Penicillium fungus on CDR discs
  2. Publication list - publication list for Dr. Cameron Jones
  3. Nanosounds - molecular computing sound remixes

I would like to ask budding molecular biologists/physicists what might actually happening here and for the programmers, "conjecture on what might happen to source code or binaries on a *fungal CD*".


Observe what happens to source code, object code or binaries when CD media they are contained on is covered with Lactobacillus bacteria.

Cut some code to CD, or use an existing CD with binary code. (Be prepared to throw the CD away). Apply bacillus to CD, culture bacullus and report results. Read and view, *carefully* - (ie: clean inner ring of CD in order for CD to be mounted) to see what happens.

The bacteria suggested is Lactobacillus . any type of CD media, computer and source/object or binary code.


  • does the bacteria smeared CD allowing the program to to execute?
  • is the programming result still the same?
  • explain what is the bacteria doing to the code and how does this effect execution?

It's a bit of a lame experiment as their are a lot of variables but has anyone had experience or willing to try this and write it up?


Journal Journal: datetime.perl.org

DateTime talk at Melbourne.pm

Had Ian Tuskett talking about DateTime modules in development at Melbourne PM last night. It was a pretty slick talk with clear overheads, examples and amusing anecdotes.

Tried some code out at home using Cygwin (DateTime installation failed), then ActiveState - also failed.

Here the ActiveState ppd collection status including DateTime - (red == failure, CPANRun "D" Builds Status (last updated Fri Nov 7 14:18:45 2003).

Another reason why ActiveState (read perl on windows) sucks. Time to ssh/telnet to my *nix boxes and try.

Slashdot Top Deals