HNHacker News
TopNewBestAskShowJobs

blipmusic

153 karma · joined January 29, 2017

submissionscomments
blipmusic··on Why We Use Julia, 10 Years Later

    > Would you be willing to provide an example where this becomes a problem?
It's not a specific example, as much as every piece of code where mutability is involved. I don't have a LOC number in mind, but it could be for blocks/local scopes. IMO syntactic explicitness helps clarity and thus bug hunting/maintenance, and I feel tooling shouldn't be an alternative for dealing with, albeit subjective, syntactic shortcomings.

    > this should be a problem for structured applications and hints at a more fundamental design problem
Frankly, I don't see how introducing an explicit hint that a variable is created can ever be a bad thing. Explicit lack of mutability is also something I'd like to have (e.g. 'let' vs 'var') - sometimes I want to make sure not to accidentally change a value. We will shoot ourselves in the foot at some point, so a syntax that allows for a check whether I can actually change a value further down can only be a good thing.

I absolutely do not trust myself. :)

When things go wrong for this reason (difficult to spot typo etc) I may waste time on investigating whether blocks are actually leaky or not etc - I have issues with Python syntax for the same reason. Going in the other direction, I like Rust's perhaps overly verbose "let mut" much better, since this has other consequences for explicitness (mutable references '&mut X' vs immutable references '&X'). No special tooling or IDE required to spot this, I can 'cat' the code in a terminal and it's still obvious.

I absolutely don't think Julia or Python should "go Rust", since their dynamic nature is a strength, especially so in academia where notebooks are popular. But a simple "I'm creating a variable" should IMO be a nice visual cue to any developer.

blipmusic··on Why We Use Julia, 10 Years Later
Please, yes!

It may sound dumb, but such a "simple" change (simple conceptually, can't speak of the dev challenge involved) change would be an enormous quality of life increase to me. I've ended up looking into scoping rules and what not when bug-hunting because of this, but often just ended up going back to a language with stricter syntax around variable creation (goes for moving away from Python etc as well). The more the code, the bigger of an issue this becomes.

blipmusic··on Archivematica: Open-source digital preservation system
Ouch, thanks for the heads up. :/ We'll be stuck with FITS this time around, though.

Also, someone should put OCFL on the front page, while HN is in an archiving mood.

blipmusic··on Archivematica: Open-source digital preservation system
I've seen experiments built on top of https://3dhop.net for viewing (if that's the issue), but have no experience myself. Don't know about animation, though.
blipmusic··on Dublin Core's dirty little secret (2010)
In terms of wasted space, only 1/8th is actual xmp for that DJI exif chunk, the rest is padding for (13mb jpg:s in our case). 8k, so 1k actual xmp data of which a lot is structural. I've only tried with scripts, verifying with exiftool and a hex editor every now and then.
blipmusic··on Dublin Core's dirty little secret (2010)
Depends on what device generates the image perhaps? DJI Drones have a lot of info as EXIF XMP. I just parse that as XML.

EDIT: It's not 80k, but good enough as a GPS log.

blipmusic··on Archivematica: Open-source digital preservation system
It seems so, unless I'm misunderstanding the documents. From an old, random readme for a workshop, then the official wiki:

    - https://github.com/mjordan/archivematicaworkshop#archivematica-and-aip-migration
    - https://wiki.archivematica.org/AIP_re-ingest
Perhaps someone with better knowledge than I could tell whether this implies that each version gets its own PID (crucial for publications etc).
blipmusic··on Archivematica: Open-source digital preservation system
Others out there, such as:

    - DataVerse: https://dataverse.org
    - Omeka: https://omeka.org/s/
    - also many internally developed archival systems
To us it's mostly up to the granularity of the metadata, and wether metadata fields support e.g. ISO standards for validation (which is why it's difficult for a single metadata standard to rule all, unless all you need is free text descriptions). Another need is a way to batch ingest.

    > Compatible with hundreds of formats
I'm becoming increasingly critical of file format verification as a gatekeeper for digital preservation, but perhaps someone can explain why I'm wrong. Striving towards open specifications is of course a must for long term preservation (cave paintings persevere, while digital data is fragile in so many ways), but if conversion incurs data loss, parallel archiving should be an option. Worse, we've had issues with old systems that deny archiving data althogether because of an automated format checker standing in the way (we'll use FITS [0] in the future, but that just uses a bunch of other type checkers under the hood and doesn't seem bomb proof from the little testing I've done). Formats such as MP4 also seem like a nightmare to validate (granted, lots of cameras out there ignore following specifications). But archiving nothing at all over a few proprietary formats every now and then is a horrible outcome (see experience above). At the very least, it must be possible to override automated format checking if necessary.

    [0]: https://projects.iq.harvard.edu/fits
blipmusic··on Dublin Core's dirty little secret (2010)
There are also modular approaches like CMDI [0] out there, that allows for constructing a metadata profile with a corresponding schema [1]. A single metadata standard won't be able to cover all disciplines regardless. Too many free text fields that can't be properly validated. E.g. if you work with documenting minority languages, language ISO codes are vital (assuming one has been created for the language in question).

    [0]: https://www.clarin.eu/content/component-metadata
    [1]: https://www.clarin.eu/content/component-registry-documentation
blipmusic··on History of the Nautilus loudspeaker
:-)

Also see Glenn Gould's Goldberg variations:

    https://www.nytimes.com/2007/03/12/arts/music/12conn.html
    https://www.prosoundnetwork.com/archives/yamaha-disklavier-and-zenph-restore-goulds-1955-igoldberg-variations
    https://www.classicstoday.com/review/review-13701/
blipmusic··on Show HN: Audino – Open-Source Audio and Speech Annotation Tool
How does this compare to ELAN (https://archive.mpi.nl/tla/elan) in regards to doing the actual annotations/transcriptions? Or could ELAN/EAF-files perhaps be considered for input formats in future releases?
blipmusic··on Hacking GitHub's Auth with Unicode's Turkish Dotless 'I'
Julia returns:

    julia> c='ı'
    'ı': Unicode U+0131 (category Ll: Letter, lowercase)
    julia> uppercase(c)
    'I': ASCII/Unicode U+0049 (category Lu: Letter, uppercase)
    julia> lowercase(uppercase(c))
    'i': ASCII/Unicode U+0069 (category Ll: Letter, lowercase)
Is this something that needs changing in the Unicode spec itself or how strings are handled in general by various tools/programming languages? I love [plain] text, but it's so, so fragile. :/
blipmusic··on GoPro Hero 7
Considering this is HN, does anyone have any input on the latest GoPros vs Garmin Virb Ultra 30, specifically in terms of telemetry formats vs general camera quality? Garmin seems to be the data king here, but I'm curious as to whether the GoPro is perhaps the better camera, and if so, whether this could offset the differences between the data formats.

I should note that the telemetry is important for our usage, GPS logs especially.

We have two Ultra 30s at work (used in humid areas, mostly at walking speed) and with GPS turned on (obligatory) and 1440p/30fps video we get 90-110mins of continuous footage before the battery is completely drained. Longer battery life would be nice and every now and then we get hiccups/glitches (looks like frames are skipped) with the setting above. Doesn't invalidate the footage in any way and it's rare, but still annoying. Could be a firmware bug, really. Audio is so-so, but that is to be expected (we haven't tried external microphones yet, since the setup needs to be as simple as possible).

Would a GoPro help with any of the above or is it a similar experience in terms of battery etc? If anything, the GoPro seems more rugged (for our needs, the Virb needs a case). What about video quality? Interface? (I did read DC Rainmaker's comparison[0] which was what made us choose the Virb, together with the fact that there's an SDK for FIT and at the time of purchase GoPro didn't have a Github repo[1])

GoPro's GPMF data is embedded within the video file, whereas Garmin's FIT[2] format is separate and is logged to whether one is recording or not as long as the camera is turned on. Both have their pros and cons, FIT seemingly being the more flexible format (it's basically a big timeline with events attached) but one also needs to identify what part of the data stream corresponds to which video session and it's not as "linear" to read as GPMF (?). In FIT every single logged coordinate is also timestamped (10Hz GPS), whereas it seems that you only get one timestamp/second for GPMF coordinates (18Hz GPS).

I wrote a little parser to extract and integrate the relevant telemetry data elsewhere and have started looking at doing the same for GPMF (I wanted to write it from scratch in a language I'm learning). Since I have no recent GoPro at hand it's been difficult finding raw samples for the GoPro, other than the GPFM GitHub repo - youtube won't suffice since I need the original video with the telemetry intact.

[0]: https://www.dcrainmaker.com/2017/03/best-action-cam-2017-gop...

[1]: https://github.com/gopro/gpmf-parser

[2]: SDK here https://www.thisisant.com

blipmusic··on The new MacBook Pros with T2 chips do indeed kernel panic randomly
Add lack of 16:10 screen to that. It's getting dumb. Is Apple buying up all laptop screens with this aspect ratio? Also, are non-Apple trackpads getting better?

These two things are the biggest issues every time I ponder moving to Linux (Windows is not an option or I'd look at the Surface alternatives). The P1 does indeed look very interesting otherwise.

blipmusic··on Inventing the Future
Firefighters getting up-to-date info and routes to emergency exits etc in large, burning buildings. A mechanic, medic, farmer... getting info on what's in front of them with suggested actions. People not focusing on their phone while on their bike...

Can't see how going from so-so handheld devices to something that is in your line of sight when needed and adapts to current context is a bad thing compared to what we have (not that I'm assuming the end products will necessarily be well made, but that goes for everything).

Luckily, there's more to this world than Pokémon.

blipmusic··on Transmit 5
I like Forklift[1] but that's more of a Finder-replacement with remote possibilities, I guess.

1: http://binarynights.com/forklift/

blipmusic··on Show HN: Add a play button next to IPA on Wikipedia
There are still languages with no writing system of their own out there (or with a borrowed script that does not necessarily fit their sounds very well). IPA is not perfect but quite relevant for research and language documentation.
blipmusic··on Show HN: Neural Japanese Transliteration
"Scold" or something similar to it was in there, so quite possibly yes. Thanks, for the heads up.
blipmusic··on Show HN: Neural Japanese Transliteration
I didn't mean to imply otherwise. My curiosity was/is with the generation of the model in the link (especially compared to current models), hence the question in my original post. There's also pitch accent that helps with disambiguation of spoken Japanese (I should know, I accidentally called a friend seaweed for a long time).
blipmusic··on Show HN: Neural Japanese Transliteration
Well, to be honest my first impression was that this is what I've been using all along. I just liked that someone experimented a bit. :) Perhaps I assumed this would end up with a different predictive model. No idea what kind of corpora have been used for the input methods already available.

Edit: your P.S. made me remember the "ma-ma-ma-ma..." mouthful the Chinese language students I studied in parallel with discovered (apparently well-known - something about a horse and a...mother?). If tones are not represented in romanised Chinese, things seem to get tricky, indeed.

blipmusic··on Show HN: Neural Japanese Transliteration
Does it act up on your desktop OS as well? I know there's an alternative from the ATOK developers for iOS (http://www.justsystems.com/jp/products/atok_ios/) but I haven't tried that one.
blipmusic··on Show HN: Neural Japanese Transliteration
Cool! So, would it be correct to say that this is in essence generates a disambiguation model for a language with a lot of homonyms due to having relatively few sounds (but also with some variation in morpheme boundaries, e.g. "an-i" vs "ani")?
blipmusic··on Tim Cook: Technology Should Serve Humanity, Not the Other Way Around
That's fair enough as long as other companies and their customers are held equally responsible. Just because we sometimes/always use linux does not mean the hardware we run it on has an equally "clean" background. Also, we buy too much stuff, meaning too much stuff is being produced... etc.
blipmusic··on Tim Cook: Technology Should Serve Humanity, Not the Other Way Around
Turn it off then? It's in the settings. Can't say I particularly like Siri either but not using it works so far.

So what's a real example of Apple making the UI worse to further their agenda (i.e. selling more stuff)?

blipmusic··on Sublime Text build 3133 now available
Even with telemetry turned off?

    "telemetry.enableTelemetry": false
blipmusic··on LibreOffice leverages Google’s OSS-Fuzz to improve quality of office suite
I had a bunch of old Word-documents from a late researcher using non-latin scripts and custom fonts that needed fixing for archiving (the only important step was to make these readable by humans). Don't remember what encoding. Current version of Word wouldn't even touch these, raising errors left and right.

In the end I changed the strings that specified the problematic/missing fonts in a hex editor (I naively assumed that would do the least damage...) to a compatible font I eventually found after some digital archaeology. LibreOffice accepted these files but didn't display them correctly (as in: according to what my eyes and brain expected - "machine readable" was in a different universe at this point). I saved the hex-edited documents in whatever version of docx LibreOffice exports to. Word accepted these, displayed them as well as could be hoped, and a little bit of modern history could be saved for future generations. Not a very technical approach, I guess.

I feel the 1990's in general did a lot of damage. It's not only Word. Competing, proprietary formats together with lacking text-encodings (and in my case custom fonts messing up character tables, due to academic/linguistic needs) has caused more information loss than the decades before and after for the data I've worked with.

Suffice to say I use plain text as much as possible (markdown, latex), as text-encoding alone is hard enough.

blipmusic··on Windows 10 S Will Not Support the Linux Subsystem
Is this reddit?
blipmusic··on Windows 10 S Will Not Support the Linux Subsystem
Do you want me to? :)

Every single time I fire up Pythonista on an iOS device I soon realise that any meaningful coding I do requires access to external data. And then I give up, until next time some years down the line.

blipmusic··on How we invented new ways to teach Japanese
It can be just as simple in Japanese as in English, and just as complex. They refer to ways of being implicit + grammatically encoded social deixis, which Japanese is good at (not exclusive to Japanese, though). Grammatically it's quite covert.

But on the whole, no. Just no. I think what's said in the article on this is bullshit. As for wa/ga, one can not necessarily replace the other, so you'd lose a few permutations on that alone. Also, there are more personal pronouns to choose from if that's the game.

I also did not know that just because speech level is grammatically encoded in Japanese, one is not allowed to generate similar meaning in other languages.

Or rather: I'm sure I can find 248,026 ways to say "I love you" in some other language as well.

What is it with Japanese that triggers this kind of ignorance and Sapir-Whorf-ism...?

blipmusic··on Moom removed from sale due to patent violation claim
https://www.spectacleapp.com but whose to say that won't disappear as well...? Software licenses are starting to feel dangerously close to "patenting the wheel".
← PreviousPage 2 of 3Next →