hallo leo blogs

[Mylio] is a fantastic Photo Management application. It's fast, has a clean UI, offers unmatched photo syncing between devices, is reasonably feature-rich and has great non-nonsense support people who help you out with technical tips down to tricks with the internal SQL data store.

I said 'reasonably feature-rich', but there are some features painfully missing; features you'd expect when coming from, say, Adobe LightRoom. One of these omissions is keyword autocomplete:

Mylio has a Keywords field (mapped perfectly to entries in the EXIF and XMP metadata), but it spears in the UI a basic text field where you type the keywords as comma-separated tags without any help about what keywords you have already used in other photos. That is pretty cumbersome when you have a growing photo collection where you want to keep a consistent and manageable collection of tags.

So I was looking for a way to add autocomplete externally. I immediately thought of Keyboard Maestro. Getting the details from a post on the Zettelkasten forum I built a little tool to help with keywords in Mylio. I created it as follows:

Generate a text file photo-tags.txt with all your Mylio keywords following this Mylio forum post.

Build the following Keyboard Maestro macro:

Then you can autocomplete tags in Mylio like this:

Go to the keyword field of a photo.

Trigger the macro (I use here the shortcut opt-cmd-/.)

Start typing a few letters of your tag, now you get a list of tags _containing these letters:

Now select fthe desired keyword from the list of complete tags.

Voilà! -- So far the macro works pretty well.

I'm using xonsh for three years now, but I still find little conundrums:

I had written a helper script which calls different command line tools with various arguments depending on some conditions. (Perfect use of xonsh!)

In this scenario I used one command, let's name it find_optimum and I had to call it with a description file problem.txt as argument and sometimes with an option --cache-results:

find_optimum --cache-results problem.txt

At other time the script had to call the command just as

find_optimum --cache-results problem.txt

My approach in the script was something like this:

extra_opt = '--cache-results' if can_cache else '' find_optimum @(extra_opt) problem.txt

This is nice and concise xonsh syntax, right? True indeed, apart from the fact that it doesn't work. From the find_optimum command I always got the mysterious error "Can't find file ''".

It took me a while to understand that in the case can_cahe=True the @-expression @('') is in subprocess mode converted to an empty, but existing argument -- and the called command find_optimum obediently gets this empty string as its own separate argument and tries to find a file with the empty string as name instead of the file problem.txt.

The solution becomes clear when you know that the empty list evaluates in the @-expression (@([])) to nothing indeed (and not the empty string). So use as fallback place holder for extra_opt the empty list:

extra_opt = '--cache-results' if can_cache else [] find_optimum @(extra_opt) problem.txt

Half by accident I came across What is 0 to the power of 0? on Youtube, an inspiring recording of a normal high school class session about the value of$0^0$. Half way through the video I thought: It would be interesting to see a graph of $x^x$ when $x$ is getting closer to $0$.

With Python's matplotlib (and seaborn on top) that's pretty easy, so I plotted $f(x) = x^x$ for $x \in [0,1]$:

This graph suggests that for $x$ approaching $0$, $f(x)$ approaches $1$. But lets plot a bit closer to 0:

Same story. And if we look very close to $0$:

Same story again. So let's propose:

$\lim\limits_{x \to 0^+} x^x = 1$

From this it makes sense to set $x^x = 0$. This is indeed how many mathematicians1 (but not all of them2) define it. Actually a good overview of the issue is given on Wikipedia.

See for example Donald Benson, The Moment Of Proof: Mathematical Epiphanies, 1999, p.29: "The consensus is to use the definition $0^0 = 1$." ↩︎

I am looking at Prolog at the moment. Brian Tate's Seven Languages in Seven Weeks inspires me. I am working through some of the exercises suggested in the book.

First off "Reversing a list". I have to admit I had to think quite hard about it, but the solution is pretty simple:

reverseList([],[]). reverseList([H1|T1], L) :- append(T2, [H1], L), reverseList(T1, T2).

And you query the predicate like this:

?- reverseList([1,2,3], R)

and the system answers:

R = [3, 2, 1] ? -

Pretty magic, heh?

PS: Just finished a "Smallest element of a list" predicate as well:

minOfList([A], A). minOfList([H|T], M) :- minOfList(T, E2), M is min(H, E2).

This week at the Monday sprint of the 10th PyCon AU in Sydney I "built" my first microcontroller "thing".

Well, thanks to Matt Trentini I was able to play with an ESP8266 on a Wemos D1 board) This tiny board can run a complete Python implementation!!!

And MicroPython was already "flashed" onto the device, so first thing was of course to get a REPL going. But when plugged in to my MacBook the ESP8266 didn't show up a serial tty. I needed to find the right driver (board specific). Once the driver was installed on my laptop I was "in".

The next challenge was to get the on-board LED lighting up. - That was easy: Everything worked the way the ESP8266 tutorial on the MicroPython site described.

And then I dove a little bit deeper in the boot-up process of MicroPython. I wrote a script (and flashed it onto the board), which made the board, when plugged into power, blink a few times then stop. Yeay!

Later I played a bit with some NeoPixel LEDs (on a so called "RGB Shield") connected to the board and I realised there's a lot more which can be done!

Yesterday at PyCon AU I gave the talk "Dash for Crime". It is about the data visualsation framework Dash by Plotly.

The room was choc full, which idincates that there is certainly interest in Dash and I got some good questions at the end.

In the last couple of weeks I felt my Tango dancing had been a bit uninspired. Over the years I have had phases like this a few times, so I new it will pass, but still - I wondered what could I do to give my dancing a bit of new life?

The idea

Well, I though: Ristrict yourself! Dane only simple steps. Going a bit more in detail, I decided, the steps of the *basico * will do:

Side steps, caminada, the cross and occasional back steps. All of this in varying cadencia and including cunitas.

But none of the following:

Ochos, giros, spins, walking outside of each other; not to think of out-of-axis moves.

The practice

Equipped with this clear idea I went to our lovely milonga El Boliche lways on a Monday night. I took the challenge and danced as a leaderas I had planned.

How did it work? Reasanble well from a challenge standpoint: I manged to avoid ochos and all the other "fancier" moves; only once I had to over turn after a side step to avoid touching another couple (teh floor was not very crowded).

The result

So far so good. But how did it feel? This is where the surprise comes: It did not feel good! The restriction did not free me up to listen more to the music and to connect better with my partner. To the contrary:

Obeying the restriction needed a fair bit of concentration. My dancing felt to me not necessarily un-musical, but certainly a bit monotonous - and to others slightly tired: A few times I asked at the end of a tanda for feedback from my tanda partner - and that my dancing felt "tired' was certainly an attribute mentioned more than once.

Hoping to connect better with my follower didn't work either: Very unexpectly I felt like I had less Less options to adapt well to my dance partner. This came as a quite unexpected insight to me: a broader repertoire of moves is not only more interesting (to myself and hopefully to the follower), but it seems to be like a richer tools kit when it is about finding ways to dance well with each other: With some people the simple non-turning steps are a bit hollow, but giros might work great. It is the ocho after the cross which works better than when the follower walks backwards out of the cross. Same for more complicated steps: With one follower the colgadas might work super well, while the more introspective simple style might us as a couple not give as much pleasure.

The simple style needs certainly stronger musical attunement of the dance couple. Yes, maybe we could even say generally: The simpler, more restricted the moves are, the more other aspects of the dance need to work well (musicality, embrace, etc.).

Going forward

Although this experiment did not hat all have the expected outcome and I would say it was in this regard a failure, I have learned quite a bit from it. Furthermore I hope that the concertation the restrictions afforded might become less with practice freeing me hopefully up to indeed dance more musically.

I have just finished my moreman tool.

Moreman assembles man pages for commands without man pages. It does so by utilising the help message most commands provide.

Great inside Emacs to view the possibly lenghly help of a commands in separate *Man* buffers.

I'm curious how the new flag ship iPhone in handling size compares to the other models. - So here a table listing relevant iPhone models in order of physical dimensions:

1

We can clearly see that the new iPhone X is noticeably bigger than iPhone 7 and 8, but smaller than the Plus models - it lays somewhat in between.

So we have it again: When an iPhone model with a completely new form factor comes out - like the switch from model 5 to 6 -, then the "default" size of the phones increase.

So the times where people wearing normal men's trousers could put their mobile unobtrusively in a trouser pocket are definitely gone...

Today I tried at the aptly named The Loading Zone an new beverage, a “Dirty Chai”: This is a coffee with some Chai mixed in.

I am in Sydney this weekend and I cycled over to the shoe maker Salvio’s Dancing Shoes in Randwick - just to find out that it is closing down:

I like their dance shoes quite a bit, so of course I had to ask why they are closing down. The answer is as expected: “People buy from the competition from overseas. It’s is cheaper…”

For me an era comes to an end. I danced in Salvio’s shoes for over 10 years! They fit my feet perfectly: Snug, but not tight.

My Salvios’s shoes are well-travelled. They saw Australia: Sydney, Melbourne, Canberra, Brisbane, Perth, Byron, Port Macquire even Wagga Wagga. And they explore the world: Rotterdam, Berlin, Essen, Wuppertal, Den Haag, Amsterdam, Zürich, Barcelona, Paris, Hongkong, Singapore, Tokyo, Guangzhou, Bangkok - I’m sure I forgot the odd city (like Wellington!).

When staying for a few days in Bangkok (of course to dance Tango!) I tried some barbecued fish at a Japanese street kitchen:

(Checkout the photo of the fish on the grill at flickr.com.)

I just listen to another interesting episode of the podcast Hello Internet. In Episode 48: Grumpy About Art the podcast authors CGP Grey and Brady Haran describe their visit to the exciting art project Dismaland in the UK. (I wish I could have gone and seen it!)

In due course of their discussion they come to the interesting topic of art and art expertise and talk about the arbitrariness of modern art. CGP Grey then issues the following statement (I hope I transcribe it here correctly):

There is no expertise to be had in the appreciation of art

I want to discuss this a bit, because I totally disagree with Grey. Appreciation of art is most definitely impacted by expertise.

But step by step. What does art appreciation start with? Well, it starts with what we see and this feeds into what we feel about the art work. The first part (seeing) depends in turn on two things:

What our senses observe

What our mind registers from these observations

The second aspect of registering is where expertise comes in: We potentially register more when we know what to look for, when we have an expertise about categories and history which the process of art creation might have influenced.

So the expert is able to see more. But is he or she able to feel more too? This depends on how we understand "more": If "more" means "bigger", expert and lay person might be on par: I don't think, while appreciating art the emotions of an expert are necessarily "bigger" than the emotions of a lay person. However the emotions of the expert might be more detailed because of the more levels of detail the expert has registered in the art work.

I do not say that without expertise there is no art appreciation possible: Of course we can appreciate art when we have no expertise (and that is what I do, because I am not an art expert)1. But I have to pay my tribute to experts who know more about art than I do: They are able to see more, thus they appreciated the art work in finer detail than I do.

Ugliness it is! As beauty is it!

The excellent podcast episode Ugliness of the Body Sphere program on Radio National from 26 September 20151 contemplates the intricacies of Ugliness:

What defines ugliness? How do we react to it? How does it set the "ugly people" apart from the rest?

Certainly a worthy piece of audio work. I will listen to it again!

I discovered a trick for pumpkin risotto: Roast the pumpkin pieces first and stir them only at the end into the risotto. This way the pumpkin pieces keep the nice roasted texture and their distinct caramelised taste.

Here the whole recipe:

Roast Pumpkin Risotto

Makes 4-6 servings

Ingredients

For roast pumpkin:

1/2 butternut pumpkin

4 tablespoons olive oil

1 tea spoon cumin powder

rosemary stalks

cracked salted and pepper

For the risotto:

1 onion

3 cups Arborio rice

6 cups of vegetable stock (preferable self-made from scratch!)

1 cup white dry white wine (or a dry sparkling wine if you rather want to accompany the risotto with glass of sparkling)

1/2 cup of grated Parmesan

a few rosemary leaves and some Parmesan to grate at the table

Directions

Preheat oven to 200° C

Take the skin of the pumpkin and cut in medium-sized cubes (ca. 1 1/2 cm edge length)

Mix the oil with cumin in a large bowl and toss in the pumpkin cubes. Mix it all well, so that all cubes are lightly coated with oil.

Place the pumpkin cubes in a flat, oil-lined casserole. Space the pieces evenly in the dish in only one layer. Tuck the rosemary stalks amongst the cubes.

Roast the pumpkin for ca. 45 min until caramelised.

Meanwhile start the risotto:

Heat up the stock in a pot on its own.

Prepare a large second pot, big enough for the whole rsotto including the pumpkin. Heat up a splash of olive oil over low-medium heat.

Dice the onion and fry it slightly in the large pot. Take care not to brown the onion pieces.

Add the Arborio rice. Fry for half a minute.

Stir in a cup of the hot stock. Keep stirring until the fluid is soaked up by the risotto.

Repeat this until all stock is used up.

Stir in the wine. Keep stirring for all together 20 min (from the first cup of stock).

Mix the grated Parmesan and the pumpkin cubes thoroughly through the risotto.

After a long wait, MoneyWiz 2 from the smart people of SilverWiz is out! MoneyWiz is a great financial software suite I'm using for quite a while now very happily. And I was looking forward to the upgrade for better reporting et al.

First thing I noticed was, when MoneyWiz 2 for iOS (App Store link) was released I hadn't been notified - A (one-time!) in-app info would have been appropriate, I think... That MoneyWiz 2 was out I realised just by accident through a twitter post. Immediately I bought it for the currently "discounted"1 price of 6.49 AUD.

Then I ran it and connected it to my MoneyWiz cloud data (the SilverWiz people call this "SYNCbits"). Everything went smoothly; MoneyWiz 2 for iOS loaded the data and I enjoy it since on my iPhone and iPad. On the iPad (an iPad 2, which is a few years old) the app runs a tad slow, but is still usable.

So far so good. But then I headed back to my Mac for some more serious budgeting - and what did (the old) MoneyWiz 1 for Mac tell me?: It cannot connect to the cloud data, because another device has upgraded the data without backwards-compatibility!

I was pretty annoyed! Firstly MoneyWiz 2 for Mac wasn't available yet2, and secondly this meant I now have to buy MoneyWiz 2 for Mac if I want to do budgeting on the Mac as well! I did intend to buy the Mac versionin any case, but it is ridiculous that SilverWiz didn't warn me more clearly before using or even better buying MoneyWiz 1 for iOS!

OK, they have an in-app warning when reading/upgrading the data in the new iOS app, but it wasn't obvious that the old Mac software would not be able to connect anymore. There should have been:

A clear in-app warning with spelled-out ramifications plus

A big info on the front page of their website

I am quite disappointed by this lack of transparency. - The SilverWiz guys do such a great job with their software, but the failure to inform us long-standing users about this issue shows a certain lack of customer-focus.

SilverWiz has to manage the discounting via time-limited price reduction, because Apple offers no other means of upgrade pricing - like many people have pointed out, this sucks! ↩︎

Age is a rather special “property” of us: Unlike hair colour or political opinions we cannot change it. At any point in time it is just given to us.