HNHacker News
TopNewBestAskShowJobs

delhanty

1,230 karma · joined April 12, 2014

https://x.com/delhanty

Moderate countabilist

Making DOG’s ModeLer (#DOGML)

dogml.eth - expiration date: 2072.09.14 at 17:07 (UTC+09:00)

submissionscomments
delhanty··on Mathics – A free, light-weight alternative to Mathematica
I'm confused.

The lead developer up until 2 years ago, Angus Griffith's LinkedIn says he's at Google. [0]

In April 2016, Angus replied to a comment of mine "Mathics is still maintained. There are currently two of us working on it. The last release was v0.9 in March this year." [1].

The original creator Jan Pöschko's GitHub profile states "JavaScripter making the notebook interface of wolframcloud.com" [2]. His bulk contributions to Mathics stopped in 2012, but he still makes sporadic contributions to Mathics - see here he merged a pull request March 2019 [3] - suggests Wolfram tolerate the existence of Mathics because it's not a serious threat.

As the creator of Sage, what's your opinion of the legal risks of creating a Mathematica syntax compatible clone like Mathics in the light of Wolfram Research's fairly aggressive assertion of its IP "rights"?

[0] https://www.linkedin.com/in/angus-griffith-093433b0/

[1] https://news.ycombinator.com/item?id=11465064

[2] https://github.com/poeschko

[3] https://github.com/mathics/Mathics/commits?author=poeschko&s...

delhanty··on Gun laws where guns take the most lives
>However, there is a loophole. In a country of nearly 132 million people, there is only one shop that sells firearms, which is located in Mexico City. ... This high level of unregistered firearms has been attributed to Mexico’s proximity with the United States, as thousands of gun retailers sit just over the border.

It sounds like it's the Mexicans that need a wall to protect themselves from U.S. firearms!

delhanty··on What happens when Google says “we aren’t going to pay your fines”?
A counter-example: https://www.gov.uk/ works pretty well. *

In fact, it's less user-hostile than most commercial sites.

Why is that?

* Disclaimer - I'm British

delhanty··on Ask HN: Highest paying remote companies?
>You need to setup presence in all the countries you have employees in.

For that reason, many non-US remote workers are technically contractors.

delhanty··on Mark Zuckerberg Speaks Down to Users and Misses the Point
>that got flagged that ended up with a ton of upvotes after it was clear it a genuine comment.

That happened to me too!

I'm happy that HN has the flagging feature, but I suspect that feature is being abused in an organized way by rings to hide comments that run counter to certain vested interests.

delhanty··on Facebook to integrate the infrastructure for WhatsApp, Instagram and Messenger
>To add insult to injury, WhatsApp permits rekeying to take place without any indication to the conversation's participants [in the default settings].

Agreed.

A nice Twitter thread [0] by Mustafa Al-Bassam back at the end of November on how that might be exploited by GCHQ:

>Ian Levy of GCHQ has released an essay on how law enforcement should get access to end-to-end encrypted communications. Here is the critical bit to pay attention to. They're proposing to exploit the fact that users don't verify each other's public keys, and inject bad keys.

Then this [1] later in the same thread by Twitter @inag_fc:

>This is a coordinated attack by 5 eyes. They slipped it through AU parliament in the week, presumably as some horse trading because there was practically no debate nor warning, beyond the normal straw man proposals.

[0] https://twitter.com/musalbas/status/1068179464197156864

[1] https://twitter.com/iang_fc/status/1071373264646225920

delhanty··on Oracle v. Google and the future of software development
Just me maybe, but I'd rather we didn't compare any people at all to cancer here.
delhanty··on Design flaw behind MacBook Pro’s “stage light” effect
From the linked Wikipedia article:

>The basic concept of MagSafe is copied from the magnetic power connectors that are part of many deep fryers and Japanese countertop cooking appliances since the early 2000s in order to avoid spilling their dangerously hot contents.[2][3][4] Apple was granted US Patent No. 7311526 on MagSafe ("Magnetic connector for electronic device", issued in 2007) as MagSafe was deemed to be a sufficient improvement due to the connector being symmetrical and reversible, and the fact that magnets within a connector are arranged in opposing polarities for improved coupling strength.

IMO, it's outrageous that improvement was enough for the patent not to have failed on prior art grounds.

I have one of those Japanese countertop cooking appliances. The magnetic connector is symmetrical. It's very obvious that Magsafe ripped off the design.

delhanty··on Ask HN: What software do you actually pay for?
>Devexpress Winforms and ASP.net component suite (best Windows component suite on the market - leaves free alternatives in the dust)

Question (for the Mac people here): what's the Mac equivalent of that?

delhanty··on Apple resumes selling iPhone SE at lower $249 price on its clearance store
Korea and China - not spent long in either, but can quite believe that larger phones would play better culturally in those countries than in Japan.

Japan - small does generally play very well here - think back to how small some of the Toshiba Libretto notebooks were.

A high-quality, small phone; easy to hold; easy to store in a small designer bag - once the novelty factor of a bigger phone has worn off - it's hard to understand why that's such a niche play.

Only Apple, Google, Samsung etc. have authoritative data. But from Apple's point of view, cancelling the iPhone SE would have simplified their supply chain, which would have been an easy sell to Tim Cook as he's an operations guy.

That doesn't mean that we'll never see another small phone from Apple again.

delhanty··on Apple resumes selling iPhone SE at lower $249 price on its clearance store
>People who want a smaller phone in 2019 are a very small yet very vocal minority.

Really?

Here in Japan, there are a lot of ladies with small hands, not particularly vocal on HN, and they seem to find small phones more comfortable to hold.

delhanty··on Ideology Impairs Sound Reasoning
>Of course antivaxxers are irrational lunatics, ...

Why is that "of course"?

It looks to me you haven't got the point of what #ReedX wrote at all:

>Remind yourself of the times you were confident about X and turned out to be wrong.

Or #haberman:

>Embrace humility, since we're always wrong sometimes.

delhanty··on Ideology Impairs Sound Reasoning
>seek the best form of the opposing viewpoint

Strongly agree with this. Let's try a specific example on HN:

Seek the best form of the view that the anti-vaccine people may have a point.

delhanty··on You Weren't Meant to Have a Boss (2008)
Because it's a transparently self-serving article perhaps?

YC combinator is in the startup business, so of course PG is going to oversell the case for startups.

PG's stuff is hit and miss. Here's an example of a much better article by him:

Don't Talk to Corp Dev

http://www.paulgraham.com/corpdev.html

delhanty··on 9999999999999999.0 – 9999999999999998.0
Exactly!

Very far from a floating point expert here, but what I do is to scale-down by a few odd prime-power factors as appropriate:

Scaling down by powers of 5 is obviously appropriate for decimals, currency etc.

Scaling down by powers of 3 is good for angles measured in the degrees, minutes, seconds system.

If one scales down a lot there is an increased risk of overflow, so one can compensate by scaling up some powers of 2.

The way I think of this is as using my own manual exponent bias [0].

>the exponent is stored in the range 1 .. 254 (0 and 255 have special meanings), and is interpreted by subtracting the bias for an 8-bit exponent (127) to get an exponent value in the range −126 .. +127.

So, for example, even single-precision number are always exact multiples of 1/(2^126), and I'm just changing the denominator to contain powers of 3, 5, 7, ... etc.

[0] https://en.wikipedia.org/wiki/Exponent_bias

delhanty··on OpenSCAD: Software for creating solid 3D CAD objects
FreeCAD is nowhere close to the quality level of Fusion.

Sketch-Up and OpenSCAD are just very different from Fusion.

The closest cloud competitor is OnShape: starts at $1500 per year - only free option is public storage.

https://www.onshape.com/cad-pricing

delhanty··on OpenSCAD: Software for creating solid 3D CAD objects
>If I could buy fusion 360 for $1000

Why would it be that cheap? Fusion 360 is competing with SolidWorks. A standalone license for SolidWorks probably starts at $4000.

For any comparable commercial 3D parametric mechanical CAD system just the B-rep modeling kernel alone represents several hundred person years of commercial development going back 30 to 40 years, and academic research going back 50 years.

The Fusion 360 kernel is ShapeManager [0], forked from ACIS 7.0 in November 2001.

Work on ACIS [1] started in 1985 after the 3 founders (Ian Braid, Charles Lang, Alan Grayer) left ShapeData having worked on Romulus.

Romulus [2] was released in 1978 and was based on BUILD 2.

BUILD 2 and before that BUILD were created by Ian Braid during his PhD thesis started in 1969 and supervised by Charles Lang. [3]

Now, what about SolidWorks?

The SolidWorks kernel is Parasolid [4], dating from 1987, which was a C dialect rewrite of Romulus, which was written in Fortran. ACIS, by the way, was written in C++, so was one of the first commercial systems written in C++.

How do I know the above? I worked on Parasolid 1989 to 1995. We had 20+ devs then - department head count was 40+. The still have lots of devs working 30 years later.

There's only one comparable open-source B-rep modeling kernel - Open CASCADE, that powers FreeCAD. But quality wise, IMO comparing it with Parasolid, ACIS or ShapeManager is a bit like comparing Lucene with Google or Bing.

CGAL that powers OpenSCAD is high quality - but it's a very different beast, and not something you could use to power Fusion etc.

Why is B-rep modeling so much work? Here are some reasons:

* B-rep models are inherently complicated graph based data-structures with many different node types (geometry and topology). Just keeping the model consistent is a lot of work.

* Lots of surface and curve types means lots of combinations for surface X surface, surface X curve, curve X curve intersection/distance/... algorithms.

* Tolerant modeling.

* Topology case analysis from hell for blending.

* Hybrid b-rep modeling with other approaches, meshes, volumetric etc. and make it work properly.

* Feature-based history modeling layered on top of b-rep modeling means that your b-rep operations must replay EXACTLY the same, even when you upgrade the kernel.

[0] https://en.wikipedia.org/wiki/ShapeManager

[1] https://en.wikipedia.org/wiki/ACIS

[2] https://en.wikipedia.org/wiki/Romulus_(b-rep_solid_modeler)

[3] http://solidmodeling.org/awards/bezier-award/i-braid-a-graye...

[4] https://en.wikipedia.org/wiki/Parasolid

delhanty··on Spain will soon overtake Japan in life expectancy rankings
"The USA is the IHME index’s biggest loser among high-income countries falling from 43rd in 2016 to 64th in 2040, due to an increase of just 1.1 year – from 78.7 to 78.9 years."

I would have made that the headline.

delhanty··on A New Largest-Known Prime Number
Read this and wondered, what estimates are there on the density of the Mersenne primes as n grows?

Google told me that, surprisingly, it's not even known that they are infinite!

I found this nice page on the Wagstaff Mersenne Conjecture:

https://primes.utm.edu/mersenne/heuristic.html

delhanty··on Facebook has always sold data to advertisers, and it probably always will
I do pay for stuff on the internet.

IMO the NY Times simply isn't good enough or original enough to be worth paying for though, so I choose not to subscribe to it.

Also, I'm not very interested in being informed about the endless news cycle of daily events. It distracts focus and in the long run is mainly noise. Its main purpose seems to be to create jobs in the media for the oversupply of liberal arts graduates.

delhanty··on Facebook has always sold data to advertisers, and it probably always will

  >Please disable your ad blocker
  >
  >Advertising helps fund Times journalism.
No I don't think so - uBlock Origin all the way.

Internet advertising fuels the Facebook problem.

The smaller we can make internet advertising, the smaller the Facebook problem becomes.

delhanty··on Why Use F#?
Thank you.

>it would certainly be possible to share 90% or more of your core code.

It sounds like maintaining parallel code bases in Ocaml and F# with the same basic design and file layout might work quite well for the core.

delhanty··on Why Use F#?
Thanks very much for that perspective and the link.

It certainly sounds more attractive than writing everything in Javascript and using Node!

I'm interested in writing a portable, commercial 3D CAD modeling library. F# has already been used commercially for some years for 3D CAD on Windows. So F#, when compared with Ocaml, is a relatively easy sell into that market as it doesn't add another dependency.

delhanty··on Why Use F#?
I'm looking at Ocaml again because of the existence (and quality) of the Bucklescript backend for Javscript.

That makes Ocaml sound pretty attractive on the portability front.

A question for anybody here regarding portability: Is it viable to maintain a portable core using a subset of Ocaml (something like Standard ML maybe - no objects) and compile that in F# using Ocaml portability mode?

For context, I've read this question on SO from 2010 [0] but I'm looking for some more 2018 perspectives from HN people.

[0] https://stackoverflow.com/questions/4239121/code-compatibili...

delhanty··on Ask HN: Who is hiring? (December 2018)
Looking for a job in Tokyo?

Sign-up for HN Tokyo Slack [0] and check-out the #_looking_for_staff channel.

NB. No professional recruitment agents allowed.

[0] https://hntokyo.io/

delhanty··on Go 2, here we come
>Generics in c++, c#, scala, java

They're not all the same though are they.

C++/CLI had both compile-time generics (templates) from C++, and run-time generics from the CLR. And they could be complementary at times.

For compile-time generics, Dlang are a lot more sane that C++, having had the benefit of coming later, and dumping C compatability.

Similarly, CLR (C#, ...) generics had the benefit of being designed having seen Java first, so IIRC they're baked into the CLR. CKR generics were derived from work done at MS Research Cambridge, and I seem to remember Don Syme (F# creator) being rather proud of them. Disclosure: I was contracting at MSR Cambridge back in 2007.

Anyway, the golang designers will be aware of these implementations, so hopefully they'll come up with a nice design.

delhanty··on Sheryl Sandberg Is Said to Have Asked Facebook Staff to Research George Soros
Furthermore, the New York Times and friends are competitors with Facebook, Google et al in the market for online advertising, so their interest in taking down Facebook comes across as self-serving.

For myself, I'd like to see a 95% contraction in the online advertising market $ value, so I'm uBlock Origin all the way.

Edit: for clarity, I'm not here to shill for Facebook - they're toxic! I'd be quite happy if Facebook shrunk to $0.

delhanty··on A note to our employees
It definitely sounds like a step in the right direction.

This is from Stephanie Hurlburt's Twitter a while back.

  Ah, game industry culture and drinking.

  I don’t drink. I also feel unsafe being in a mostly-male group when they’re all drinking.
  
  ...
https://twitter.com/sehurlburt/status/999707171381100544
delhanty··on Why Are Antiques So Cheap? Because Everyone Lives in the Kitchen
Just for future reference for me or anyone else on HN:

I found the George III inlaid mahogany bureau referenced in the story on the Gorringes (Lewes, UK) website [0] and saved to archive.org/is [1][2].

The sale price was £70, and the pre-sale estimate was £100-150. The story referenced a sale price of £85 including fees, so I guess the fees must have been £15.

[0] Lot 794, A George III inlaid mahogany bureau W.97cm https://auction.gorringes.co.uk/auction-lot-detail/170918/79...

[1] https://web.archive.org/web/20181104033127/https://auction.g...

[2] http://archive.is/mkxdn

delhanty··on Do we really live longer than our ancestors?
OT - sorry

>On the other hand, because I live in rural Japan and am only really friends with the people who live near me, almost all of my friends are over 70 (I'm 50).

I'm in a pretty similar situation ... guess from your profile that it's OK to hook up with you by email ?

← PreviousPage 6 of 15Next →