Trial period for CBD-Zoo direct bus

Metlink published a press release today to the effect that they will be introducing a direct service (numbered 23z) between the CBD and the Zoo, starting from Feb 3 (a Sunday) as a six-month trial. Rather than following the route of the old #10, it will run down Adelaide Road.

There is currently no timetable information available from the website, but we do know it will be seven days a week. Hopefully it runs reasonably frequently, so that tourists can just show up to Courtenay Pl and get on without having to find a timetable. The old #10 ran half-hourly; on the other hand, it doesn’t seem efficient to run this new service more frequently, because the #23 through the area is already half-hourly and so clever timetabling should lead to an every-15-minute service between Newtown and the Zoo

As well as providing direct service to the zoo from the CBD, the service will provide some extra capacity on Adelaide Road to Newtown.

One also hopes that some promotion will take place beyond the short-term: the #18 in Auckland is loudly advertised in the paper timetables as being “the tourist bus to the Zoo and MOTAT”, and that (anecdotally, at least) seems to translate well into tourist usage.

What I’m trying to say, I suppose, is that Metlink should stick a large photo of a giraffe onto the bus timetable.


By the way, let me just complain about something. The #18e has been running hourly Mon-Fri since October, so it should count as a standard route, not a peak route; but the timetable still shows it as grey rather than blue:


Further, it’s listed here as an “extended peak route”, rather than just an extended route!  I will be watching the new #23 timetable with interest to see whether this mistake is made again (although there will still be extended peak routes on there, the #23e and #29e, so I would expect the wording to be something like “standard, extended, and peak routes”).

On the topic of numbering, here is a quick map showing all the #23’s which will be running after Feb 3:

Brown: #23, Mon-Sun 2BPH; Blue: #23e, Mon-Fri peak; Green: #23z, Mon-Sun ?BPH

I am a little worried that three different routes with basically the same number but totally different service patterns might be a bit confusing; on the other hand, Wellington City bus numbering has always been a bit odd (after all, there are two buses numbered #30x that go to totally different ends of the Miramar peninsula)!

Finally, Metlink has been on a pamphlet-publishing spree recently (it’s interesting to note that two years ago their brand presence was practically non-existent – if the bus was late, the average person would blame GOWellington or VF – and now they’re ubiquitous – every problem is now their fault!), with info sheets on bike racks, and on how to use double-decker buses correctly (and finally a new stand-alone paper network map, but unfortunately a Wellington City-only one), so perhaps we’ll see a tourist-focused pamphlet to go with the old Gold Card- and child-focused pamphlets.

Edit. A photograph from the railway station, showing some Metlink advertising for services to the zoo.


What I want to see in Wellington transport this year

Now that the new network is up and running, I want to write down what I think the next steps should be.

Improved frequent services. The #7 service only runs half-hourly on Sundays. The same applies to the #110, #120, #130; and the #220 runs even less frequently, operating in a 20-40 pattern on Saturdays and Sundays. I would like to see all frequent services (except the #21 and #22) operate frequently seven days a week, and the addition of a frequent service in Kapiti (probably the 260). I don’t think any other non-frequent services deserve to be upgraded at this point (although the 160/170 pair could be better timetabled to provide a proper 4BPH service to Wainui between them).

Better advertising for routes to tourist destinations. I would like to see destinations like the zoo, Zealandia, and the regional parks better emphasised on paper timetables and on buses – especially as we see tourist numbers increase.

Better visibility for the ferry services. I would like to see more advertising, as well as better signage and inter-mode ticketing options, for the Day’s Bay Ferry. In addition, this year it would be nice see the reintroduction of the ferry service to Petone (especially if the third ferry we keep hearing rumours about gets purchased).

More electric buses. The GWRC is looking at introducing electric buses around the region outside Wellington; they have also been in talks with NZ Bus about keeping the battery-converted former trolley buses in the city. Let’s hope we don’t see any stalling, and that both of these go ahead this year.

A LGWM report. It’s now about six months late. I hope it includes funding for light rail down towards Newtown; more realistically, I would expect to see at least:

  • A second Mt Vic tunnel (I think this is non-negotiable for NZTA).
  • Some kind of at-grade redesign for the Basin.
  • Improved cycleways and pedestrian facilities on the Quays.
  • Improvements to Adelaide Road (going beyond the Newtown Connections cycleway project to deal with traffic volumes and bus priority).

I also wouldn’t be surprised if NZTA wants to trench Karo Drive, but I’m not sure if they could dual it to take traffic off Vivian St (the bottleneck is the tunnel under the war memorial, which is three lanes and no hard shoulders – both the motorway trench under Ghuznee St and the road reserve around Karo Drive are wide enough for four or more lanes in most places).

Wider consultation on the Regional Land Transport Plan review. Consultation with govt agencies on a new draft PT plan was expected to start last month, and wider public consultation is scheduled for April. (Source: agenda for the 5 Dec 2018 meeting of the GWRC sustainable transport committee, from page 50). I would expect to see more in this about the possible purchase of electric/diesel multiple units for service to Masterton (and maybe even Palmy); this is even more likely after the government last year gave GWRC extra funding to upgrade the Wairarapa line.

Construction on the Hutt Valley-Wellington cycle link. I know it’s not scheduled to start until 2021, but I hope the next budget pumps in more funding to get it started sooner. We should see the start of a detailed design next year, though.

Improved services to Otaki. With the opening of the M2PP expressway, and the Otaki bypass and Transmission Gully scheduled to open in 2020 (next year!), GWRC needs to start improving public transport in Otaki in preparation for a population boom. The land is cheap and psychologically it’s about to get much closer to Wellington (despite the actual time savings of Transmission Gully being small and the fact it’s likely to increase traffic volumes), so it would be nice to see PT improved from the infrequent 290 bus to something more substantial, if only to offset the extra demand caused by the roads. In the same vein, we should start to see bus services in east Porirua increase as well over the next few years (given that the offramps from Transmission Gully will open up heaps of new land, and the Kiwibuild investment in the area) – but I think Otaki is a more important target in the shorter-term because the infrastructure for growth is already there. Even before we get new DMUs for the Kapiti line, I would like to see a shuttle service of some kind between Levin and Paraparaumu (remembering that Waikanae is hard to terminate trains at from the north, given it has only one platform). The old MAXX units are still parked up at Taumarunui, and electrification to Pukekohe will free up the DMUs – they’ll do in the interim.

Improved services to Masterton. We already know Metlink has ambitions for longer trains to Masterton, with a 9-car train expected into service sometime in the next few months. There are also track improvements planned, including maintenance work on the WRL north of Upper Hutt, and double tracking between TREN and UPPE.

Design for the rail corridor cycleway in Upper Hutt. The last annual report said that “The preliminary design was completed and provided to KiwiRail. Work is on hold until
KiwiRail is able to dual track the railway line.” Let’s see this get started. Further safety upgrades to the Upper Hutt portion of SH2 (the Hutt Expressway, Moonshine Bridge, and River Road) would be welcome.

I think that’s about all I have to say; I’m looking forward to the opening of the upgraded Evans Bay cycleway and the Kilbirnie/Miramar cycling upgrades; and until LGWM publishes their report, I don’t have anything more to say about public transport.

Happy new year!

Easy topology exercise

At the end of a semester: this post is dedicated to the other people who studied in Ned’s Cafe but to whom I never spoke.

I want to prove the following, which is an easy result that either appears in Munkres as a theorem or as an exercise.

Theorem. If X and Y are topological spaces, and if Y is compact, then the projection mapping \pi_X : X \times Y \to X is a closed map (that is, if F \subseteq X \times Y is  closed, then \pi_X(F) is closed).

For the proof, we will need the following tube lemma (whose main application is the proof that the finite product of compact spaces is compact):

Lemma (tube lemma). Let X and Y are topological spaces, and let Y is compact. Let x \in X . If A \subseteq X \times Y is an open set such that x \times Y \subseteq A , then there exists some U \subseteq X such that x \times Y \subseteq U \times Y \subseteq A .

Proof of lemma. For each (x,y) \in x \times Y , let V_y \subseteq A be some neigbourhood of (x,y) . The family \{V_y\}_{y \in Y} covers x \times Y ; since the latter is homeomorphic to Y , it is compact and so there exists some finite subfamily \{V_i\}_{i =1}^n \subseteq \{V_y\}_{y \in Y} covering x \times Y . Let U = \cap_{i = 1}^n \{\pi_X(V_i)\} (which is clearly open); I claim that this is the required U . Indeed, we have x \times Y \subseteq U \times Y by construction. Further, if a \in U \times Y , then \pi_X(a) \in U and so \pi_X(a) \in \pi_X(V_i) for all i . Then a \in \pi_X(a) \times  Y \subseteq \cup V_i \subseteq A . This proves the lemma.

Proof of theorem. Let F \subseteq X \times Y be a closed set. Consider some x \in (\pi_X(F))^\complement . Then x\times Y \subseteq F^\complement ; by the tube lemma, there exists some open U \subseteq X such that x \in U and U \times X \subseteq F^\complement . Hence x \in U \subseteq \pi_X(F^\complement) = (\pi_X(F))^\complement , and so each point of (\pi_X(F))^\complement has a neighbourhood entirely within the set; so (\pi_X(F))^\complement is open and so \pi_X(F) is closed.

Things which annoy me, #7243

Introspection week continues today, sorry…

I was browsing stack exchange and I was reading a question on the front page; here’s a quote (I apologise to the asker of the question in advance, I’m not trying to persecute them in particular):

Can I reject the premise of a question in an interview?

This actually came as a practice (not marked) question in a video interview for an internship at a bank, where I was asked to name my “favorite movie” and explain why.

My actual response to the question (in my head) was “I’m not sure what the word favorite means — I don’t know of an operationalisation of favoriteness in a linearly ordered set with an injective function from the set of movies — but here’s a movie I like: _____ …”

Instead, I simply gave an expected answer.

Now I don’t care about the question per se (well, I cared enough to read it) – but I do want to briefly talk about this quote because it is a good example of something which really annoys me, and isn’t the New Year just the time for hating things.

In case you haven’t spotted it, or in case you don’t know me so you don’t know how annoying I find this, it is this gratuitous usage of basic mathematical and/or scientific terminology. I have no idea why people do it; here’s another vague example from the TV show that I could probably write another blog post on if I cared (in fact, I know there are better examples of this phenomenon in that show, but my brain is already in pain from finding that one – broadly speaking, my relevant complaints to this post are talked about in this video, although that doesn’t necessarily mean I agree entirely with that that video is saying). This thing also has cropped up a lot in the recent Math with Bad Drawings posts; that’s not even a blog I dislike (I think it’s even in the sidebar on here, and also half the time the examples on there are actually ironic). Incidentally, remind me someone to write the post I keep meaning to about this MwBD post.

I don’t know why people feel the need to do this: are you trying to sound smart? (In which case everyone who knows how to use the words you use think you sound like you’re just trying to sound smart, and everyone else just thinks you’re an arsehole.) Do you actually think that using these words make you more understandable? (Surely not.)

Mathematical terminology, like ‘injective function’, is used in order to make statements precise and clear. One will say ‘the function is injective’ in lieu of saying ‘the function is such that inputs can be uniquely identified by outputs’, because it shortens and/or clarifies mathematical prose (it encapsulates a single, well-used concept); but that doesn’t mean that you can just drop it into random sentences. Sure, it makes sense in the sentence above – but it’s pointless, because it’s simpler and more clear in this situation to say ‘I don’t think there’s an objective way to talk about films being better or worse’, or even (if you really want to push some kind of mathematical metaphor) ‘I don’t think there’s a well-ordering of films based on favouriteness’.

And that’s even putting aside the fact that what the asker here is saying is nonsensical to begin with: “an operationalisation of favoriteness in a linearly ordered set”? What does that even mean? I think that what they mean is that they have some set which contains values which quantify favouriteness, and they have placed a well-ordering on it. Whatever they mean, it’s certainly not clear to me – and if I have to guess what they mean because of imprecise language, then their use of mathematical terminology is arguably pointless. (I have looked up the meaning of ‘operationalisation’ and it does seem that this is what they meant, but it’s still a mixed metaphor at best and a giant stretch at worst.)

So my point is that if you want to use mathematical language, step back and ask yourself two questions:

  1. Am I just doing it because I’ve just learned these cool new words and I want to show off?
  2. Is it easier and clearer to understand what I’m saying if I don’t use a mixture of statistical terms and naive set theory to describe it?

I don’t think I’m asking too much: after all, there are plenty of situations in which using naive set theory and statistics drastically increases the clarity of statements.

A job interview in which you are being asked your favourite film is not one of them.

Just more generally, I want to include in this category (because I’m apparently categorising things which annoy me) those science podcasts which unnecessarily bring in terminology from elementary physics or calculus for no real reason; it just feels like they’re introducing an artificial barrier for entry – if you’re not talking about sophisticated ideas, you shouldn’t need sophisticated language. Not that I’m against people trying to talk about elementary calculus or physics in podcasts, it’s just the unnecessary use of words which I object to. I think the video I linked to above about The Big Bang Theory is somewhat relevant here too: the problem I have is people referencing the terminology of something on a surface level (replacing the word “acceleration” with “second derivative of displacement” for no good reason, for example – I’m aware this is a fairly tame example) in order to artificially raise the perceived sophistication of their content, or to ‘tag’ their content as being ‘sciency’.

Why is this important? The main reason I care about this so much is that people seem to fall for it. I think it would be great if we could, as a society, be in a place where people could listen to and create scientific content in a thoughtful way, rather than just throwing around some buzz-words and thinking that that’s enough (again, I should stress that I’m not singling out anyone here – nor could I, even if I wanted to. I think everyone is somewhat guilty of this, and it’s something a lot of people are aware of anyway – I’m not spouting new ideas). It’s fine if what you’re saying makes sense, but the problem is that people who want to go out and spread ‘fake science’ (I wish there was a better phrase) can just throw buzz-words around as well; and if everyone is just throwing words around as ‘tags’ for their audience rather than to actually further an argument or clarify a point, then there’s no way for the general public to discern an actual scientific argument which cites replicable data from a reputable peer-reviewed study from pseudoscience.

(<sarcasm>Join me next time, when I for no real reason yell at people who talk about calculus like it’s the most sophisticated pinnacle of mathematical achievement.</sarcasm>)

Why I don’t study computer science

Apparently it’s introspection week here on the bloaaeeg, so if you’re here for, like, content, please come back sometime next week. (We want to do some more Public Transport 2050 stuff, we’re up to Porirua from memory, so look forward to that once we’re on holiday.)

Anyway, it’s story time.

In the before times, I was very interested in computer science. I learned to program in primary school, and got really interested in functional programming (with Haskell) and real object orientated programming (with Smalltalk-80, I can’t remember which implementation but it was the one installed with Scratch 1.4) in secondary school. I used to have a really weird desktop installation, I think it was Arch with the Openbox WM. I used to contribute to some open source projects (if you Google my name, at least one of them still shows up). I remember when KDE 4 and Gnome 3 were released, and I remember when Ubuntu switched from Gnome 2 to Unity, and I remember struggling with SysV init scripts and hating SystemD because it’s just not UNIX-y. I used to have an opinion on the editor wars. I was originally planning to study computer science and physics at university when I finished college.

So what happened?

First of  all, I did a bunch of things in Y13 that convinced me to study mathematics: I did an overseas summer school in Bulgaria, and studied first-year mathematics via the MATH 199 distance course from UC.

Secondly, I dropped DGT in Y13; this was mainly because the New Zealand digital technology curriculum was awful (it’s currently being reviewed by NZQA, I haven’t looked at the new standards), and we had a little group which did our own thing at lunchtimes anyway. Also, I’ve never been that interested in efficiency and studying algorithms theoretically and stuff like that (I can write doubly-linked lists, or stacks, or whatever, and I use them all the time, but I’ve never used a slow enough system or large enough dataset to care about speed).

And thirdly, perhaps the main reason I chose to study mathematics instead of compsci rather than as well as it, is that in computer science nothing ever works. This is actually what inspired this post: I’m currently running a weird hybrid of Manjaro and Arch on my laptop with KDE5, and I did an update yesterday and rebooted today and now can’t find the Qt5 shared libraries (I’m getting messages like ‘/usr/lib/ not found’, even though they’re there, point to the right version, and have the right checksum, and even after running ldconfig). I remember I used to have random problems like LibreOffice dialogs growing the bookmarks pane over time on their own (I think this actually ended up being a bug in Gtk and got fixed earlier this year, I can’t immediately find the post on Phoronix though), and random applications just don’t work (I’ve always had a lot of trouble with VLC and I have no idea why), and file indexers like Baloo never work, and using the KDE GUI to power off the system locks somewhere in the Plymouth splash screen, and everything is always so fiddly and it all integrates together like someone ran out of rivets and had to use duct tape.

I should clarify here, it’s the implementations that I have (philosophical?) problems with. The actual designs and structures are generally incredibly well-designed; I think Haskell, and really Lisp, the greatest programming language of all time (that song was, like, my JAM, dude, in Y12), or in fact functional languages in general, are beautiful. I really love the idea of objects living in space, and everything is done via message-passing, and no-one knows how anything is implemented. I think there was a Joel on Software post (wow, I haven’t been on there for ages – he’s changed his theme, it annoys me now) about how theoretical compsci graduates spend all their time designing these edifaces of design, with interfaces on top of interfaces, and abstract classes inheriting from abstract classes, and reflection and encapsulation, and are incapable of actually writing something to do anything useful at all; which sounds about right to me.

Maybe I stopped being really interested in programming when I started using Python: it’s so utilitarian, and as a language I never found it expressive or fun to use. It’s impossible to shoot yourself in the foot, there’s a library function for literally everything so it all feels like an assembly line, and I hate languages where indentation is meaningful to the compiler. (I used to use a lot of C++, can you tell?) I like being either close to the metal or powerful abstractions, and I suppose Python is just too practical. (Python is to Smalltalk as calculus is to topology, and C++ is to Python as the Peano axioms are to calculus; and assembler is to C++ as ZFC is to the Peano axioms. And real programmers use gamma rays to flip bits.) I used to be interested in writing small kernels and running them in qemu, but I never really got into compiler design for some reason.

(In addition, I am incredibly worried about ethics: the subject is new enough that a lot of concrete work is being done in all kinds of directions, but the vast majority is done without reference to privacy, security, or ethical implications. In this sense, it reminds me a lot of economics.)

I am reminded of the passage from one of Feynman’s books (I forget which one) in which he describes the feeling of burning out from physics, except I ended up doing something else which I seem to enjoy more.

As I think I mentioned in an earlier post this week, I’m planning to do some work in Smalltalk during the summer, so that should be fun (one hopes).

I haven’t mentioned physics (and the reason I’m doing physics is perhaps more interesting than the reason I’m doing mathematics), but I’m tired of nostalgia blogging… if there’s any interest, I’ll write that post next year.

Addendum. I’ve managed to begin to fix the issue with the missing Qt5 libraries, I’m still not entirely sure what’s going  on (I’m on Manjaro unstable, and it’s broken – surprise surprise).  After some fiddling around, I found that the package which installed this file – qt5-base – was at a weird version (5.12.0.r312.ge80bf655e9 rather than 5.12.0) that Repology tells me is outdated. I checked further, and this is indeed the version sitting in the Manjaro unstable repo. However, the qt5-base package is also provided by qt5-base-old-statx which seems to be the correct version (5.12.0) – removing qt5-base and replacing it with this package (running ‘pacman -Syyu qt5-base-old-statx’ uninstalls qt5-base at the same time because they are listed as incompatible; doing it like this also avoids having to force pacman to uninstall dependencies) seems to have fixed the problem, and KDE is now running for me. This is clearly a temporary solution, but I was planning to reinstall my system around Christmas anyway (likely moving back to Debian unstable). So PSA: don’t upgrade your Manjaro Unstable systems right now!

This didn’t fix Firefox (which was also broken), but simply clearing my package cache and reinstalling firefox seems to have worked.


I’ve been seeing a lot of stuff around on social media and elsewhere recently, and I’m so sick of it that I’m genuinely angry.

Being racist, sexist, or homophobic is not OK.

It’s not about political correctness. People aren’t telling you to shut up because they’re being PC, they’re telling you to shut up because you’re an arsehole.


I don’t care who you’re talking about, whether it’s the prime minister or someone working in the checkout at the supermarket, keep your opinions about them as a person to yourself. Stand back and ask yourself if you would be happy for people you don’t know to be yelling that stuff at you (either in real life or over the internet), or openly judging you based on how you appear, especially if it happens every day for most of your life.

Free speech says that you can’t be locked up for having opinions, but it doesn’t say that anyone has to listen to them or that they can’t tell you to go shove your [redacted] into a [redacted].

If people tell you they’re offended by something, don’t keep saying it. It’s not “standing up for free speech” or “fighting political correctness”, it’s just being an arsehole. I seriously don’t understand how some of these people can say stuff on social media, it’s attached to their real name and on a public page, and they seem to think that it’s their right to be seen to say these things.

A lot of the discrimination I’ve been seeing in real life recently hasn’t even been conscious; it’s things like being “macho” when talking about maths (follow the link, it explains what I mean better than my wording here), or talented and motivated female high-school mathematics students (linked post tangentially related, but blog highly recommended) going on to study biology or medicine or the soft sciences because they don’t know any girls going on to study physics or mathematics at university (I have had a couple of students for whom I knew this was the main reason for choosing their particular field of study at university), or just flat refusing to believe there is a problem at all.

Seriously, we should all work to be better than this.