- Free Software - Comments

Things have to change

South Bruce Peninsular (trackback) — Sun, 05 Dec 2004 00:31:42 -0500

Things have got to change.We can argue and we can throw figures at each other..and we can hire other people to be hired guns, but the best is to talk..discuss..with NO-ONE that has another motive.No hidden agendas..nothing but what is right, and what c

OpenContent ... or BitTorrent?

garym — Tue, 17 Aug 2004 17:44:17 -0400

Although this page has had tens of thousands of page views, I do wonder: is it useful?

Keeping this page in sync with the releases takes some effort to monitor the release directories and then hand-edit the Open Content Network links -- that's a lot of bother if most people who arrive here either don't have Java (so their download reverts to plain HTTP) or if they just use the BitTorrent (most Linux distros configure Mozilla to automatically handle BT links).

So what say you?

* I should keep the Open Content Network links even if they occasionally get out of sync, at least they let people know the list of mirrors where they can find the ISOs, and besides, the BitTorrent maintainers are very often weeks behind the release dates.
* or should I just keep links to all known BitTorrent hosts and a few HTTP mirrors and let people sort it out on their own.
* or did you just find this page was frustrating because it wasn't what you thought it was and leaves you no better off than when you arrived!

Comments?

Lost in Translation

garym — Tue, 17 Aug 2004 17:17:37 -0400

bq. Check this box if you only edit system files to configure your Servers

Can anyone translate that from the literal French? In some cases, it counts as one of _Select only N answers_ but in other cases it seems it's a modifier; I've applied for the position as Mandrakesoft's English Interpreter, and some of the odd twists of this survey do underline their need for some native language experience.

The survey is also not asking me about what I change, only if I change specific things, and then not ask what it is I change about them. For example, in several editions of Mandrake I had to change my keyboard to accept the ALT key as the Emacs Meta, and I changed that by editing text files because I get tired hunting through tedious hierarchical menus systems ...

By the looks of the questions, they've decided to trim the fat from their menues, although there's no indication that they might consider such bold moves as only putting _one_ text editor on the board of a menu pop-up we have to look at every single day or ease the navigation customizations so we can quickly anneal _and retain through upgrades_ an effective list of just what we need to do what we do.

Other OOo Print Resources

garym — Fri, 30 Jul 2004 09:52:03 -0400

Some other recommended OpenOffice.org print resources only very slightly out of date and still very useful:

OpenOffice.org 1.0 Resource Kit ("website":http://www.getopenoffice.org/main.html)
By: Solveig Haugland and Floyd Jones
2003/Sun Microsystems, Inc.
OOoSwitch ("website":http://www.hentzenwerke.com/)
By: Tamar E. Granor
2003/Hentzenwerke Publishing
OpenOffice.org for Dummies
by Gurdy Leete, Ellen Finkelstein, Mary Leete
2003/For Dummies
also available in PDF ebook format
OpenOffice.org ClueSheets
2002/Publishing Power

O/S Addiction Research Foundation

garym — Fri, 09 Jul 2004 14:04:20 -0400

Hmmm ... that's really quite apropos the more I think of it: Nobody defends an addictive substance as much as those either addicted to it, or those addicted to the cashflow it brings them, and the more highly addictive, the more paralyzingly total and frightening the prospects of breaking the vicious cycles, the more vehemenently the addict will defend and protest those who would detox them.

Sadly, though, as we all know, it only takes one psychiatrist to change ta light bulb, but the light bulb has to _want_ to change, and it likewise doesn't occur to most addicts there is even a problem, whether of substances, of destructive behaviours and beliefs, or the deadly combination of the two that we see in ICT vendor-addictions. It's not until, as happened with France, as happened with the agencies in nearly all these spectacular switch-over stories, it isn't until they hit the wall, until they bottom out, until they nearly suffocate and have no where else to go, and _that's_ when we can finally talk to them about IT freedom.

Kind-a sad, true, but the way of the world nonetheless; the alternative is kidnapping CIOs for rigourous forced detox cult-deprogramming, and I don't see much chance of that happening :)

Still, maybe there's an opportunity in here for a new slogan for GNU ...

bq. Switch to GNU: The _Apomorphine_ of Operating Systems

Lest we get too smug ...

garym — Fri, 09 Jul 2004 00:21:49 -0400

and before you run off and nab a Mozilla product thinking the mere swap of one surface application and all is well with world again, you'd do well to note the "news reports of a similar gaping hole in Mozilla and Firefox":http://slashdot.org/article.pl?sid=04/07/08/2159244&n where the real kicker of the commentary on SlashDot puts their finger on the exact pulse of the problem:

bq. "Note that this only affects users of Mozilla and Firefox on Windows XP or Windows 2000."

_badda bing badda boom_ they got ya all over again.

Not "Horsing" Around

Tue, 06 Jul 2004 22:35:19 -0400

I guess you can now safely say that organizations that make that leap of faith to use StarOffice are truly not "horsing around".

Check here for a really good ad running in Japan.

Microsoft can roll out countless studies that point out how proprietary software is cheaper and more practical than open source code. But how can Microsoft counter open source software's ability to make a Japanese woman give birth to a horse in a convenience store?

Quick, what do Microsoft and Big Tobacco have in common ...

garym — Mon, 07 Jun 2004 20:19:51 -0400

This is just getting too cool too fast, be still my beating heart: "Disinfopedia has been asking some obvious questions":http://www.disinfopedia.org/wiki.phtml?title=AdTI-Funding ... like, just for instance, who is it who pays all those whopping salaries at AdTI? Well, gee now, if you want to fling dung, hey, there's a game _everyone_ can play ...

bq. In June 2002 open-source software advocates wondered if an AdTI criticising open source software was actually a veiled Microsoft response to recent reports of rising government and military interest in open-source systems. Wired magazine reported that a Microsoft spokesman confirmed that they funded the Alexis de Tocqueville Institution. "We support a diverse array of public policy organizations with which we share a common interest or public policy agenda such as the de Tocqueville Institution," the spokesman wrote in an e-mail."

An unspecified spokesman, an email, I love it.

More? You want _more_? Of course "there's more":http://www.disinfopedia.org/wiki.phtml?title=AdTI-Funding ...

how quickly we forget

Wed, 02 Jun 2004 16:38:42 -0400

Strange how much interest this generated as a prospect and how few people ever actually cared when NASA began officially releasing Open Source Software about 3 months ago.

There's not much there but it looks like a start.

http://opensource.arc.nasa.gov/

About time...

Wed, 19 May 2004 02:01:48 -0400

They started getting the kids off drugs!

BC
Edmonton
Calgary
Regina
AB

Locating the Online Component

garym — Sun, 09 May 2004 01:17:02 -0400

Ah, well, _that_ was web-friendly! To get the skinny on how to join up with the online component, you have to _download the PDF file_ where it is indeed spelled out quite plainly in the sidebar of the 3-column print brochure:

bq. The conference will be webcast live worldwide and made available after the event in archive form using the KMDI ePresence interactive webcasting system http://epresence.kmdi.toronto.edu/

oh ... did they mention? Registration for the _webcast_ is $95 -- that should trim out the riff-raff, eh? Little wonder they left that gem for the fine-print footnotes of the PDF.

Ok, on second thought, I think I'll wait for the post-mortem blog distillations ... and maybe spend the $95 I don't have on mom.

Webcasted

garym — Sun, 09 May 2004 01:00:57 -0400

No hints as of midnight Saturday as to where and how this thing is going to be 'webcasted', and I note with some resignation that this 'comprehensive' program makes no mention of the community or _labour saving device and not product_ aspects that are the true meaning of free software -- hopefully this will surface and all we can do is revert to pre-dotbomb methodologies of polling the page over and over to find out what's going to be what with it.

You can watch for page updates here on the communiqué in the SiteCloud sidebar, especially as there's no blog, and no RSS feed, which I suppose is unsurprising considering this is a venue who see fit to _host_ their webbycasting on a "Windows Server 2003 running Microsoft-IIS/6.0":http://uptime.netcraft.com/up/graph?mode_u=off&mode_w=on&site=http://osconf.kmdi.utoronto.ca/

Soured defaults in the cdrecord

garym — Mon, 19 Apr 2004 01:33:57 -0400

This is one where I went ahead and bugged Warly way too premature: I'd sought an answer for the name of my CD writer instead of the _application_ where the symptom occurred, and when I return to the Bugzilla to look for cdrecord, on search on the qa.mandrakesoft.com and I got some good hints on a solution pretty fast:

bq. "the defaults don't work on my system" is actually "the defaults don't work on quite a few systems," and they did work in 9.2 and earlier.

Somebody broke something with respect to detecting ATAPI-style CD burners, and that means all applications which call the cdrecord utility must now specify the _type_ of the device along with its SCSI device numbers.

So, for example, my old dev=0,0,0 option won't cut it with 10.0, I have to specify the full dev=ATAPI:0,0,0 and then all is well with the CD burning, the device is found, understood and set into motion just like back in 9.2 -- I still don't know why my FlashCard reader seems to prevent further ripping of WAV files off audio CDs, but that can be left for another day ...

A Case Study: Basing a Business on Free Software

garym — Tue, 13 Apr 2004 12:17:02 -0400

Apropos to this thread, here's a transcript of a discussion over accountability and support issues -- the scenario is the discovery that a certain piece of core-business software, free software, behaves differently on different platforms.

The issue is who they should hold accountable for the failure, and issue not far from the indemnification issue, and with the same answer: It's our code, so it's our problem ...

This client is very open minded about free software, but our conversation shows how some of the old expections can surface:

11:10 p this should behave the same across versions
11:11 p but it could be client needs to install different versions or whatever
11:11 garym not across version, but certainly across platforms, only it is under no obligation to behave the same.
11:11 p yes, across platforms
11:12 garym don't forget: Free software is the responsibility of the user ... if it doesn't work for you, you have to change it yourself -- it's a labour-saving device, not a product that has someone responsible for a contract on it.
11:12 p right
11:12 garym and that's a risk in using free software, if you approach it like it was a product. it is only there to save us from having to write it ourselves.
11:13 garym that said, the authors are most often interested in issues like this.
11:13 garym so it's worth while to find the name of the author in the two packages and let them know we've found a "possible portability conflict"
11:13 p so what is J's (or our) responsibility in packaging this stuff?
11:14 garym if we give out guarantees, then it's our responsibility. The code we give out is held out as ours, and the buck stops here, officially. authors are generally very helpful, but they have no obligations to their code, they just put it up for people to share it.
11:15 p yeah, i realize that. so we should go in and make the fix if necessary?
11:15 garym so, for example, we take that ticket app, which was broken, and A fixed it. We don't owe the author anything, but if we really got stuck, he'd probably answer our emails. -- if we make a fix the author doesn't like, then we've split the project and that then means we must forever maintain our version. this is why opensource people are so religious about standards.
11:16 garym it's because we're lazy :)
11:16 garym that, of course, assumes we can go in and fix it. It's likely a whopper of a program.
11:17 p and what about this one?
11:17 garym it may be possible to hire the author to fix it. but as I said, reporting the bug is a good first step because many times the authors do care and also have the time available (or make time). If they can't or don't want to, the next step is to ask them if there's someone we can pay to make a customized change.
11:17 garym same with this one, same even with RedHat Linux :)
11:17 p a whopper? can J do the fix if needed? C++ right?
11:18 garym couldn't tell you. this is C++/C but I've never looked at the code and since we have two different behaviours, I wonder if the windows and the unix are the same code base.
11:18 p but I think J should make the diagnosis too. Don't know if I'm qualified. I see the symptom
11:18 p ok, so there are options
11:19 garym yes, lots of options, but we have to maintain the perspective about the code we use. explain that we've run into what appears to be a portability issue and are contacting the authors for assistance.
11:21 garym software authors generally say "upgrade to the latest version" as the first diagnostic; that can be because they don't bother to maintain the old versions -- they only do what they need, anything more is a gift -- here again, we 'own' the code, and unlike commercial software, the responsibility is ours: if you need the old version updates, you have to do it yourself.

I don't know how this particular issue will unfold. The software authors are both coding to spec, but the spec is ambiguous on this one tiny issue, and it seems the unix crew took the first clause of the OR conjuntion, the windows crew took the second. If we're lucky, someone thought to make the behaviour selectable -- worst case, as with HTML in browsers, we'll just need to tailor the app to the platform services and put it down as just a fact of life.

That said, in my two decades using free software the cases where the authors don't jump right in to help are a tiny minority; most authors are kind, open and generous -- that may be why they chose to release their code as free software in the first place. I won't be in the least surprised if one or both of the windows and unix crews jump on our bug report and dig right in to correct this little anomaly, they may even keep us in the loop with updates or test-versions as they hunt for an amiable solution.

A House of Gifts

I've been told there is a Japanese proverb

bq. where a house is filled with Rights
there is no room for gifts.

All that earnest involvement and action by the package authors may happen, but it is not required to happen. It is a gift. As users of free software, our expectation must be that this code we use only saves us the trouble of typing it all in; the author owes us nothing.

We may not claim authorship, we may not even really understand how it works, but the free software we use is nonetheless our responsibility. Anything more is a gift.

And to our delight, the free software bubbles over with a bounty of boundless generosity.

Desktop Adventures

garym — Tue, 06 Apr 2004 09:38:39 -0400

The desktop system, a Dell L700CXE, seemed to go very smoothly through the install, and this time it did request the second and third disks (also requested a non-existant fourth) rolling right on down to the final remove the CD and reboot prompt without incident.

On rebooting, there were a few odd messages about a failure in the firewire OHCI driver loading (there is the ieee driver in lsmod which suggests that it did load, but I haven't had a chance to verify that yet), though nothing that really threw any alarms.

Not until I tried to load Mozilla with my local intranet page; Mozill a starts, but won't leave the default static file page, and won't say why. I try the intranet IP, same result (ie no result, still on the welcome-to-Mandrake page) so I try localhost and I get my page.

Sure enough, /sbin/ifconfig shows there is the lo device, but no eth0 ethernet. The 3c59x driver, for whatever reason, did not load.

the un-config

What else do you call a configuration process that doesn't change anything? I loaded up the Mandrake system config GUI page, navigate to hardware, from there to network hardware, and from there, I just follow the defaults right through the configuration process, each of them prefilled from the detection to what it should be. Hit the final OK on the last form and sure enough, we have a network now.

It was later that I realized I also needed to manually restart some of the network services ... like when May's computer wouldn't go online because the DHCP server is missing (ooops).

Up and running, but ...

Gnome 2.4 was running, all my menu configs and panel configs seemed to have weathered the upgrade, all seemed well and good and an amazingly fast upgrade, probably the fastest and smoothest ever ...

Then I tried using Mozilla.

First glitch, the usual drop-down type-ahead list off the location bar includes a link to search for the current location-box contents, and this link was visible, but did not respond to any mouse events; I later learned that this configuration setting had been lost, somehow set to use Jeeves, which doesn't exist anymore, and thus Google, knowing that we are all busy professionals, doesn't want to bother us with such trivial details, so it does nothing -- I set the option back to a Google default, and the feature came back.

The second problem was far more serious; it's now got me spooked. If I use Ctrl-N to open a new browser window, Mozilla will lock. Solid. No pane repaints, no scroll-bars, no key controlls, no window-decoration controls (except window-move and resize). It's zombified, only killable with a command-line kill. And then it gets really strange ...

Subsequent runs of Mozilla will lock on most form-edit operations. Highlight the location and press Del, type into the location, type into any page form, any of the hot-keys, and it's locked. I tried Mozilla 1.7, same, locked. Mozilla 1.5, my own built-from-sources 1.6, even old 1.4, all same, all locked.

Panic ensued.

I ran Konqueror, and lo there arose a saviour where I could reach the "Mandrake quality assurance website":http://qa.mandrakesoft.com but could find no exact match on this bug, just an old list, likely from Mandrake 8.x, on how some hotkeys were dysfunctional. I appended what I knew of this bug, just in case, because you never know.

Returning to the desktop, I gave Mozilla another shot and found that, after running Konqueror, whatever it was that was twisted into a knot in Gnome 2.4 was now untied -- Mozilla was back, the location bar could be edited and deleted, forms could be filled out

and I won't go near Ctrl-N ... or Ctrl-T, just in case, at least, not until I'm ready to shut down my browser and lose all currently open pages, which, in my daily work, is practically never.

post upgrade stress syndrom (PUSS)

So that's the tale so far: The desktop is running, and performance under Linux 2.6 does seem faster (XEmacs, for example, now churns my 300-msg inbox list in a quarter the time) but I now live in fear of rebooting. If it goes down, will it all come back up? I'll have to face that fear, someday, but for right now, for today, there's other stuff to do.

Like using the thing.