HNHacker News
TopNewBestAskShowJobs

nanis

3,002 karma · joined December 13, 2014

Submissions/comments/upvotes are not endorsements. Opinions my own; they do not represent current, past, or future employers/clients etc.
submissionscomments
nanis··on Unicode 15.0 Slide Show
Yet, for example, it distinguishes upper and lower case versions iota[1] from "upper case i" and "lower case dotless i". A similar provision could have been made for upper and lower case versions of "Turkish dotless i" and "Turkish dotted i". But it wasn't.

That means "ι".upper().lower() produces "ι" but "ı".upper().lower() produces "i" (unless you do it on a computer set to a Turkish locale:

    >>> "ι".upper().lower()
    'ι'
    >>> "ı".upper().lower()
    'i'
Note we have:

    LATIN CAPITAL LETTER IOTA
    LATIN SMALL LETTER IOTA
    GREEK CAPITAL LETTER IOTA
    GREEK SMALL LETTER IOTA
    MODIFIER LETTER SMALL IOTA 
    TURNED GREEK SMALL LETTER IOTA
    CYRILLIC CAPITAL LETTER IOTA
    CYRILLIC SMALL LETTER IOTA 
    APL FUNCTIONAL SYMBOL IOTA
    MATHEMATICAL CAPITAL IOTA 
    MATHEMATICAL BOLD SMALL IOTA 
    MATHEMATICAL ITALIC CAPITAL IOTA 
    MATHEMATICAL ITALIC SMALL IOTA 
    MATHEMATICAL BOLD ITALIC CAPITAL IOTA 
    MATHEMATICAL BOLD ITALIC SMALL IOTA 
    MATHEMATICAL SANS-SERIF BOLD CAPITAL IOTA 
    MATHEMATICAL SANS-SERIF BOLD ITALIC CAPITAL IOTA 
    MATHEMATICAL SANS-SERIF BOLD ITALIC SMALL IOTA
That's lotta iotas[1]. There's more[2]:

    GREEK SMALL LETTER IOTA WITH DASIA
    GREEK SMALL LETTER IOTA WITH PSILI AND VARIA
    GREEK SMALL LETTER IOTA WITH DASIA AND VARIA
    ...
The point is that almost all of these variations on the theme with explicit duplicates get their upper and lower case distinct codepoints, yet "lower case Turkish dotted i" needs to for some reason map to 0x69 in ASCII and "upper case Turkish dotless i" needs to map to 0x49 ASCII.

It is stupid and directly contradicts the statement:

> The Unicode Standard provides a unique number for every character, no matter what platform, device, application or language.[3]

The statement on the Unicode.org web site cannot be attributed to ignorance.

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

[2]: https://www.unicode.org/charts/PDF/U1F00.pdf

[3]: https://unicode.org/standard/WhatIsUnicode.html

nanis··on Unicode 15.0 Slide Show
No, the fact that there is no codepoint that makes those mappings ambiguous is due to the way Unicode decided to save to codepoints for seemingly no good reason.

What should be _the_ value of `"I".lower()`? Or, "i".upper()?

And please don't bring up locales. The whole point of accepting the complexity of Unicode is to be able to take a document which stands on its own without external references.

> Early character encodings also conflicted with one another. That is, two encodings could use the same number for two different characters, or use different numbers for the same character.

> The Unicode Standard provides a unique number for every character, no matter what platform, device, application or language.[1]

Those statements are outright lies: Unicode does not provide a unique number fpr "upper case Turkish dotless i". Nor does it provide one for "lower case Turkish dotted i".

If it did, it would be possible to correctly map "i" to "I" or "İ" and "I" to "i" or "ı" without having to know anything other than the source codepoint.

The font does not even come into play here.

[1]: https://unicode.org/standard/WhatIsUnicode.html

nanis··on Unicode 15.0 Slide Show
And yet there is still no unambiguous lower case "I" or upper case "i".
nanis··on Cobalt is suddenly abundant
Could you list some?
nanis··on Lights have been on at this school for a year because no one can turn them off
> Too balance this perspective, let's remember that not all public workers are bad people.

"This perspective" said nothing about "public" workers being bad and only referred to the implications of the incentives they have at scale.

nanis··on So why a News Feed?
> that is not a public good paid collectively via taxes.

Keeping in mind that most things which are paid for in this way do not meet the economic definition of a "public good" and therefore could have been provided efficiently by private entities for profit.

"News" is different.

At its core, without external regulations, "information" is not rivalrous: If I know something, that does not preclude you from knowing it. If I make a decision based on that information, that does not preclude you from making other decisions using that information (in contrast, if I eat a particular apple, you cannot eat the same apple).

Further, to the chagrin of many, "information", at its core, is not excludable in that there is always a way in which it can be transmitted which has frustrated and continues to frustrate many who seek to control it.

"News" is different though. Because "news" is usually something that demands attention now. Human attention is a constrained resource both because time is limited and because of cognitive constraints. Therefore, if one wants to understand what to do with this piece of "news", one may not be ability to process the this other one. Further, one can be excluded from receiving "news": Sure, you may get to find out eventually, but only people who pay for X can get the "news" immediately in a digestable way etc.

When private entities try to provide information people value, there will be a variety of them, competing, contradicting etc.

When a non-accountable body is charged with providing information that all must receive with funds collected in a way that is unrelated to what people want from the information they receive, guess what will happen.

nanis··on Lights have been on at this school for a year because no one can turn them off
> Public bodies rushing to accept contracts thinking that suppliers have their best interests in mind is horribly naive.

"Public bodies" do not rush ... People who work there do. Those people really do not have that much of an incentive to be careful. They do not really bear the costs of their decision. They are also people, susceptible to having their decision influenced by whatever small or not so small immediate benefit they can derive regardless of the ultimate cost to the people who actually pay for the consequences of their decisions. And, this is before one brings in the possibility of corrupt collusion between the people who spend other people's money and people who receive it.

nanis··on Formalising a new proof that the square root of two is irrational
We do not know what was in the tweet simply because it is not referenced/cited.
nanis··on Formalising a new proof that the square root of two is irrational
The tweet is the "secondary source" whereas the Tumbler post is the "primary source" in this context[1]:

> For example, suppose you are reading an article by Brown (2014) that cites information from an article by Snow (1982) that you would like to include in your essay. For the reference list, you will only make a citation for the secondary source (Brown). You do not put in a citation for the primary source (Snow) in the reference list. For the in-text citation, you identify the primary source (Snow) and then write "as cited in" the secondary source (Brown).

Regardless of one's preferred style guide, it is odd not to credit the author whose work that led to one's discovery of the inspiration for one's own work.

[1]: https://guides.library.uwm.edu/c.php?g=1007936&p=7303662

nanis··on Formalising a new proof that the square root of two is irrational
> Recently, somebody shared on Twitter a new proof

As far as I can tell there is no link to this tweet in this post which is bad form to say the list. The post has multiple links promoting author's other posts, but not to the raison d'etre for this post.

Clarification: The post does link to a Tumbler post which one might infer had been mentioned in the tweet, but without a link to said tweet, it's odd. It is common to give credit to the work that led to one's discovery of related work.

PS: Apparently, the author is Director of Research, Computer Laboratory, University of Cambridge which makes this lack of proper citation even more perplexing.

PPS: Compare the intro paragraph of this post to one that properly cites related work[1] with actual links to the referenced works:

>> Yesterday I came a across a new (new to me, that is) proof of the irrationality of \sqrt{2}. I found it in the paper “Irrationality From The Book,” by Steven J. Miller, David Montague, which was recently posted to arXiv.org.

PPPS: It is especially ironic that the author has also wrote this at some point[2]:

>>> Many people write their related work section in a wholly passive mockery of English, never mentioning the names of their colleagues: ... It’s discourteous and reads badly. Reference numbers are not names, so why do you keep forcing your reader to look them up in the bibliography? The direct way is also clearer: ...

[1]: https://divisbyzero.com/2009/10/06/tennenbaums-proof-of-the-...

[2]: https://www.cl.cam.ac.uk/~lp15/Pages/Scream.html

nanis··on Internal document explains why Google has become slow and bureaucratic
Apart from the intra-company dynamics, there is also the fact that Google do not face the same kind of consumer driven market pressure as, say, Walmart.

I do recommend the excellent "Allocation under Dictatorship: Research in Stalin’s Archives"[1] where the greatest incentive individuals face in an organization have involve doing relatively better compared to other individuals in the organization without significant outside indicators of the right direction.

In that sense, Google is like a government behemoth. Most involved in its direction and functioning have incentives that are decoupled from whether the people who are touched by their "products" and "services" are happy with them.

[1]: https://wrap.warwick.ac.uk/164/1/WRAP_Harrison_jel05.pdf

nanis··on Slack's private GitHub code repositories stolen over holidays
> BleepingComputer further observed that the "meta" tag containing the "noindex" attribute was itself placed towards the bottom within the page's HTML code, in an elongated line that overflows without breaking.

Placing meta tags at the bottom, out of sight ... an _interesting_ choice given the nature of the news.

nanis··on Ask HN: Has anyone found a way to avoid Windows Update KB5012170
I do have a 809 MB recovery partition (side effect of going from 80 GB HD to 2 TB), but I believe this wants a UEFI partition (which does not exist because this is a BIOS boot only machine).

On the other hand, since the recovery partition seems to be unnecessarily large, maybe I can tinker with that. That's an idea.

nanis··on Wing programming language: A cloud-oriented programming language
Hmmmm ... So, Terraform redux with a JavaScript facade?

The associated blog post[1] makes a case for infra-structure as code and not much of a case for this particular intermediary. On the other hand, given that it seems to me most uses of Terraform are wrong, a more opinionated approach might be beneficial.

The difficulty with, say, giving an app an S3 bucket is not in whether you run `new cloud.Bucket()` or something else ... The difficulty is in correctly configuring it for the current use case while enabling adaptation to future use cases. This is the kind of thing not solvable at the language level (people creating the thing need to understand the implications of the decisions and how to update implementations as conditions change). After that, they can write organization or project specific object storage implementations and only create buckets through that implementation. If the requirements fit with the object storage facility of all cloud providers, one can even make this all cloud-agnostic so that the consumer of the object storage creation facility need not worry about cloud specific details.

But, none of the things that matter are known to a language designer.

I use and advocate for Terraform not because it is great, but because it gives me a decent solver and acceptable uniformity in declarative syntax.

[1]: https://medium.com/@hackingonstuff/cloud-why-so-difficult-%E...

nanis··on I/O is no longer the bottleneck
Related posts from my past experience from about 10 years ago:

* Splitting long lines is slow[1]

* Can Parallel::ForkManager speed up a seemingly IO bound task?[2]

In both cases, Perl is the language used (with a little C thrown in for [1]), but they are in a similar vein to the topic of this post. In [1], I show that the slowness in processing large files line by line is not due to I/O, but due to the amount of work done by code. In [2], a seemingly I/O bound task is sped up by throwing more CPU at it.

[1]: https://www.nu42.com/2013/02/splitting-long-lines-is-slow.ht...

[2]: https://www.nu42.com/2012/04/can-parallelforkmanager-speed-u...

nanis··on Ask HN: Has anyone found a way to avoid Windows Update KB5012170
It turns out this is a tool for preventing Windows from upgrading from one release to another. My goal is to prevent Windows from attempting to install very specific security update[1] as it fails because it is not relevant to this particular system.

As Microsoft says:

> When attempting to install this update, it might fail to install, and you might receive Error 0x800f0922. ... We are presently investigating and will provide an update in an upcoming release.

If I can figure out how to block the update from being applied, I would be happy. I do not want to try to figure out partition moves to create a system reserved partition etc.

[1]: https://support.microsoft.com/en-us/topic/kb5012170-security...

nanis··on Ask HN: Has anyone found a way to avoid Windows Update KB5012170
Thank you. I will look into it.
nanis··on The 12-bit rainbow palette
Hmmm, with 12 bits, I was expecting 4096 colors, not 12.
nanis··on How does Windows decide whether your computer has full Internet access?
You can use your own domain and string as well[1].

[1]: https://news.ycombinator.com/item?id=28512149

nanis··on Advanced Scientific Data Format
> (netCDF, HDF) those are not "obscure" formats and are infinitesimally better than text.

so, not much better?

nanis··on Eating horsemeat in France
> Those sweet economists. (I’ve got my MSc in Economics ;)

> Giffen and Veblen goods?

Let me know when you find one. Grazing grounds for the perpetually counter-example starved.

> I’m all for the theoretical models

Nothing theoretical about what I said: Simple logic.

Your claim is that Dutch butchers would prefer to leave free money on the table?

nanis··on Eating horsemeat in France
> So, demand for lamb spikes around Easter every year. But, lamb prices actually drop around Easter (at least for consumers)

So, you don't think everyone involved in the chain of getting that lamb to you, anticipating the shift in the demand curve that occurs very predictably, don't ensure that they have the highest possible supply during that time?

You forgot the "keeping everything else constant" and changes in quantity demanded in response to just the price of the good versus changes in the shape and location of the market demand and supply curves.

nanis··on Eating horsemeat in France
> If demand is highly inelastic - as would be expected for something like culinary preferences

Demand curves, regardless of elasticity, are always downward sloping: Higher prices lead to fewer units demanded.

Your statement implies control over prices: Firms with market power do not operate along the inelastic portion of a demand curve. This is a logical implication of profit maximization.

Before one can understand that, one must internalize the fact that as a seller, you cannot choose the price and the quantity sold independently. If you pick a price, you can sell the quantity demanded at that price. If it is too high, can't sell any. If you pick quantity, you can only sell all at a price people are willing to pay.

Assume a firm has market power (can pick price) and is operating along the inelastic portion of the demand curve.

If it charges a higher price, it loses some sales, but total revenue increases (that's what demand being inelastic means -- quantity sold falls, but percentagewise not as much as you increased the price, therefore, revenue, price x quantity increases). In addition, because cost is increasing in quantity (offering more for sale costs you more) and now you are producing less, costs go down. Therefore, profit, which is revenue minus cost, must increase. Therefore, if you were operating along the inelastic portion of the demand curve, you could not have been maximizing profits.

Therefore, firms with market power (can pick price) can only operate along the elastic portion of the demand curve.

If you are claiming demand for horse meat is a vertical straight line (which is not a thing outside of being a limiting case in econ 101), then lower demand means demand curves closer to x axis origin, i.e., price is determined simply by the supply curve. With a given supply curve (which are always positively sloped due to the fact that cost is increasing in output), lower demand means lower equilibrium price,

> then price actually has to increase when demand is low.

Demand falling cannot cause higher prices (keeping everything else constant).

nanis··on Eating horsemeat in France
Horses are companion animals.
nanis··on Nonprofit boards are weird
> You can't operate a non-profit the same technique as a for-profit.

Keep in mind that profit maximization subject to a budget constraint and cost minimization subject to an output constraint are equivalent problems. A non-profit ought to minimize the resources expended to achieve a given "mission" or maximize the amount of "mission" it can achieve for a given amount of resources (depends on the circumstance), but achieving those does require it to act like a profit-maximizer.

In reality, non-profits are promotional vehicles and power exercises for people behind them with "mission" being secondary. And, obviously, people who are involved in non-profits do personally profit (reputation, influence, networking -- and in some instances by steering non-profits to spend money on things that benefit their other directly profit making activities).

nanis··on Terraform should have remained stateless
How does that work when terraforming Artifactory?
nanis··on Erlang/OTP 25.0 Release
Does no one proof-read any more?

> New PRNG added to the rand module, for fast pseudo-random numers.

nanis··on Mapr: Native cross platform maps in Rust using WebGPU
Not to be confused with MapR https://en.wikipedia.org/wiki/MapR
nanis··on Barbary Pirates and English Slaves (2017)
From the very same Wikipedia article:

> The label "Turkish" does not refer to Turkey; at the time it was a general term for all Muslims in the Mediterranean region since the majority were a part of the Ottoman Empire. During the 17th century, the majority of those called "Turks" in Algeria, were disowned Christians that had converted to Islam. They were mostly Spanish, Italians, and provençaux (French).

In my experience, most "Western" understanding of European history evolved in the last millennium is completely bereft of the recognition that it was influenced by a heavyweight to their East.

Why did Columbus sail west? Why were the Portuguese going around the horn of Africa to get to Asia?

nanis··on Barbary Pirates and English Slaves (2017)
The Mediterranean between the 1500s and the 1900s is a very interesting place with a lot of interesting shades of gray. For example: https://en.wikipedia.org/wiki/Sinan_Reis

> Sinan sailed under the famed Ottoman admiral Hayreddin Barbarossa at the 1538 Battle of Preveza against Charles' Imperial fleet and its commander, Andrea Doria. Sinan suggested landing troops at Actium on the Gulf of Arta near Preveza, an idea which Barbarossa initially opposed, but which later proved to be important for securing the Ottoman victory.

> Around 1540, Sinan's son was traveling by sea to meet him after one of Sinan's victories. The boy was taken captive by Emperor Charles' forces and was ultimately handed over to the Lord of Elba, who baptized him and raised him at court. Barbarossa made several unsuccessful attempts to ransom Sinan's son. While sailing nearby in 1544, Barbarossa sent an envoy to Elba to again attempt to free the boy. The island's Lord replied that his "religious scruples forbade him to surrender a baptized Christian to an infidel". Infuriated, Barbarossa landed men at Piombino, sacked the town, and blew up the fort, after which the ruler agreed to release his "boy-favorite".

← PreviousPage 2 of 34Next →