Wow, I hope not. As much as I am actually a Ruby fan at heart; and as much as I appreciate the R community and everything R has done, it always seems much easier to write slow and/or memory-intensive R code than in Python. Perhaps I never quite spent enough time with it but there are many corners to the language which seem unnecessarily tedious. And no references - variables are all copied around the place, which is expensive. I know, I know... worrying about pass-by-value and efficiency of assignment statements (well, R doesn't really have statements; everything-is-an-expression) means I'm doing it wrong, but most code I debug is written by someone else who is also doing it wrong..

Then there's pandas and the rest of the SciPy stack, which is the only reason I used Python over Ruby (I had also considered Perl+Moose) in my last project. pandas is extremely fast, and I was able to write some quite advanced data processing stuff which would normally have needed far more effort in Ruby or Perl.

Thanks, it's embarassing when you let a rant out on the internets... bad day perhaps. I appreciate your advice to participate in bug-reporting and forums, however these bugs and discussions already exist.

FWIW, Gnome 2.x nicely remembered the resolution and rotation settings for monitors even if it was weeks since you last had said monitor connected (via dock or otherwise). I'm given to understand that this feature was limited and made some assumptions that didn't always work out for 100% of users, but for the other 99% of us it was bliss (and I didn't realize how important that feature was until I lost it in KDE). Apparently the KDE folk want to do this thing "properly", and perhaps I should volunteer into yet another project-that-nearly-worksforme - but it still feels like my Linux desktop is going backwards.

The problem with visiting an S/FTP/CIFS/etc addresses in dolphin is that it has nothing that *looks* like an address bar. There is one, but you have to know to click the non-editable "breadcrumb-path" to make it editable. Which is more clicking... and hiding of critical functionality by way of forcing you to un-learn decades of honestly very straight-forward rules and expectations of UI interaction: editbale things should look editable! Secret-magic-cheatcodes were supposed to be frowned upon.. I always thought this was grotesquely obvious stuff - not many browsers hide their address bar, the most important input area in the application, but then again I'm sure google will fix that one day, too.

That said, I am vaguely aware that some distros mangle KDE defaults and others try improve them. Perhaps Debian is lacking here and I shouldn't blame everything on the KDE developers.

And finally, I didn't mention that I have a whole lot of respect for the state and architecture/APIs of KDE these days, it has come a long way.

No, that's the thing. Most Linux users bitch about the fact they've had to settle or compromise on the least-worst DE they could get productive with. I know scarcely few who are enthusiastic advocates for their chosen DE (except perhaps a two or three using awesome or KDE).

I never got along with KDE: been using Gnome since 1.x. I use KDE these days, despite its total utter lack of monitor management (*every* time I dock it forgets how to set my screens), and why would it EVER be appropriate to show only FOUR THINGS in the alt-tab list? Yes, I changed it, but this is a worthless default on my 12" notebook let alone for the pair of 24" screens on my desk. And I despise the K-menu (MORE clicking, BIGGER icons)... obviously, clicking your way to an app is just not the done thing: that's too hard now! Things aren't that simple any more! You're supposed to google your own machine, but do you think the file browser would let you do the same for accessing network resources? No! Somebody thought that the staggering infinity of the internet and countless permutations of local network resources at your fingertips MUST DEFINITELY be clickable from the "Network" location. Manually type a network address? Abhorrent idea! Even if that's what we do in web browsers all day, and even, god forbid, nautilus.

Despite all this, KDE is the least-worst. Well, I preferred XFCE but it had some quirks and limitations with notebook stuff.

startx -- :1

Assuming you're happy to use the locally connected display, and your local X server isn't running on :1. startx -- makes all the opts on the right-hand-side of the -- get passed on to X binary verbatim. See man X

I used JFS on all my machines from around 2007-2011, including laptops. I had many unclean shutdowns (especially on laptops) and JFS rarely had any problems, except that one time briefly in 2009 where I did actually lose a bunch of data, but then so did my ext4 reinstall a few weeks later (bad hardware).

JFS was much, much better than ext3. Especially in low-CPU situations/hardware.

I can't remember why I went back to ext4, I guess I wanted to see if it still sucked compared to JFS. With noatime I decided I couldn't tell the difference except perhaps for some really big git checkouts, but I didn't do any proper timing.

I never thought I'd own a brand new car; they depreciate so quickly. Recently my '99 Falcon has developed some expensive problems, and my TCO (including purchase/finance, insurance, tyres, servicing, rego, roadside-assist, 30,000km/year with 10% p.a. fuel cost increases) analysis showed $48-$53k over 5 years for ~$9,000 2nd-hand car scenarios, vs $56k for the Hyundai i30 diesel manual wagon I just bought for $20k AUD. If I sell the i30 for $3-$8k, I can break even (which shouldn't be hard, the run-out model I got was discounted $7.5k from list price).

As much as I loved my Falcon, on LPG fuel it cost over $8k to run last year (half of that was LPG fuel, and we normally do 40,000km/year). Almost $900 for registration costs, and we averaged ~$1,000 at each 10,000km service ($300 if nothing is wrong, up to $1500 when new parts are involved). Last time it was new ball joint on the front-left wheel, but in reality all the front suspension needs doing (new bushes/bearings, etc). The final nail is that the LPG mixer has failed so it only works properly on unleaded fuel, which costs ~50% more to run.

I didn't realize I could get a brand new family car for $20k (okay, the i30 is a lot smaller than the Falcon, but it still works for us). With half (or more) savings in fuel, and the km that we do, the option to buy new - with capped servicing costs and 5 year warranty - actually was hard to go past a new car. I think second-hand is more rewarding if you're doing fewer km and/or can maintain them well yourself without a mechanic. Both my brothers are mechanics and my dad taught me a lot about engines as a kid, but these days I just don't have the tools or the time.

I did E-mail DataHand, and they eventually responded saying they still hadn't settled on a new manufacturer (apparently their previous manufacturer no longer supports the tooling they were using; new tooling sounds expensive).

So I recently acquired a DataHand II Personal off eBay, and after a few days I'm just back up to 40-ish WPM @ 95+% accuracy for prose (I type ~90 WPM accurately on a regular keyboard). So I'm happy with my initial progress, but bracing myself for many months of practice until I can start coding with it and not get frustrated.

If I wasn't working two jobs I'd love to embark on replacing the electronics, to change the way the modifier keys work and implement bluetooth. There are already similar projects though, on deskthority and geekhack.

As for replacement schedule: cheap membrane keyboards such as the MS Natural 4000 are probably only good for a year (the keys go all mushy), but traditionally I persist with them for ~2 years which is when keys start getting stuck, some keys start requiring a lot of pressure to register, or aren't travelling smoothly. Although my brand new MS Natural 4000 I got a few months ago came with a space key that feels already buggered, out of the box (doesn't depress smoothly, binds easily unless you hit the middle exactly)

I believe it was CSIR back then, which is probably for the best because CSIROAC just wouldn't have the same ring to it :-)

I actually read about CSIRAC in highschool, at which time I thought it'd be fun to use as an online handle.

Funnily enough I now also work at CSIRO, joining 9 years after I'd already adopted the csirac handle. I've really appreciated the creative freedoms I've had, which has resulted in very productive tangents that we've developed as open source. This would not have been possible without all the great people around us enabling this kind of environment.

I doubt he's referring to income tax - "every time money changes hands" must surely be a reference to the GST.

But even if we are talking about income tax, I find it a strange complaint; I guess there are countries out there who don't have an income tax, but I'm not sure who they are. And surely things like the medicare levy and friends are a more blatant (second, third, fourth...) slug pretending to be something more noble than they really are ("It's not a tax hike, it's a levy!" fallacy).

If I had the energy to get upset about the tax system, it would be towards the double-dipping and hidden fees involved with saving and accessing and using retirement funds (and I'm ~40 years from retirement). So for a country who is trying to make retirement pensions a thing of the past, government sure does send mixed signals there by having your cake and eating the baking paper too.

GST in Australia doesn't work that way. Businesses claim back the GST paid on goods and services that went into producing goods and services they supply to their own customers. So only the end-customer who can't claim the GST component as an imput credit, actually pays GST.

So a hard disk might be re-sold 3 or 4 times before it ends up in somebody's computer, but the ATO only earns the GST once. All the other transactions are refunded as input tax credits. See Page 25 of the BAS Workbook

Nice troll. Funny that you picked Amiga, because this architecture failed to meet release criteria and was dropped from Etch in 2007.

Debian's stated goal is to be the "universal" operating system. They support any hardware that has active developers for it within the Debian project. At the same time, they don't allow a Debian release to be blocked by some architecture which lacks sufficient active developers.

Everybody admires Debian's packaging system. It's not really apt or dpkg though, that is awesome - it's Debian's packaging policies and their strict enforcement. If Debian is slow to release sometimes, it's because they're packaging and fixing so much buggy open source software, not the diversity of the architectures they support.

And this goes especially true for Samba - as GPL3 is worded, you are not allowed to use-it to serve protected files, or, if you do, it's fair game to anyone to steal your data as this whole "domain authentication" stuff is Digital Rights Management. So the lawyers say, and management will listen to the lawyers and not the engineers.

This is utter FUD, no competent legal team can come to that conlcusion, unless they're so computer illeterate they are unable to tell the difference between source code versus data or signals/transactions which might be handled by a running instance of that source code.

For example, please point to me where in the GPLv3 I do not have the freedom to write my own DRM implementation (the next iTunes? eBook reader?) and then release the source under GPLv3.

My last (indirect) experience with our legal team and funding open source development was long and arduous, taking many months but not being able protect data (the entire point of the majority of our work) was never in question.

I mean, GNU would be breaking their own GPG, for crying out loud.

We used MongoDB as a query/cache accelerator for semi-structured data. The key bottleneck was delegating queries outside of application (pre-filtering results according to ACLs, date transforms, etc.).

We don't have a shockingly huge dataset, and site traffic wouldn't be considered as webscale, but the ad-hoc schema and ability to delegate complex queries to the DBs as JS was really powerful and bought us a lot of performance for very little effort.

And it's only a cache of the authoritative data store, so we can trash mongo and re-load the whole dataset in a few hours.

Through pandas, for a start. The SciPy/NumPy stack is quite nifty, I'm especially interested in how to apply it for working with irregular time series data.

Not to say anybody should ditch R, I still support our researchers most weeks at work in using it. But it's not as clear-cut as you seem to think it is, especially in terms of memory efficiency.