Live data from Hacker News

Is my vision that bad? No, it's just a bug in Apple's Calculator

martin.wojtczyk.de

381–390 of 408 posts

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#381
My initial thoughts were that these issues were showing how things look on a non-retina display, or a display with non-default scaling. That does seem to be the case, but not entirely.

I took some screenshots and I do not see the misaligned numbers at retina or non-retina resolutions, but I do see the odd bevelled edges on the 8/10/16 "tabs": https://imgur.com/a/PqqkWai

Apple have pretty much given up on making things look correct on non-retina displays, so many things are positioned at what turn out to be half pixel steps. Depending on whether we're talking fonts or shapes things can jump by a pixel or become blurry. I wrote about this here: https://blog.gingerbeardman.com/2024/01/25/running-modern-ma...

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#383
post #305

Earlier quoted context omitted.

Along these same lines, the tabletification of Mac OS is annoying. A friend asked me to help with importing photos from the Apple Photo app on his brand new desktop Mac. The sequence of events was: Lightroom Legacy needs photos imported because the new Lightroom (cloud/subscription version I believe) has a different workflow, interface and apparently, features, so he's using both for the time being. So he follows gui…

See, the file system is a fine system for general data, but if you have data of a specific kind, then there’s often a better way than just dumping them in the file system. That’s always been Apple’s approach: let data assigned to a specific app be handled by that app [1]. Apple’s approach has also been to allow export of that data into standard interoperable formats (be it music, photos, emails, contacts, calendars,…

There has always been a vision in computing where you can access the same data with different tools.

In the Kernigan and Plauger Software Tools book that describes the Unix user space you could use tools like wc, awk, sort, uniq, and grep, bound together with the shell, to do all kinds of things on plain text files.

As a photographer of course I want to share images between Lightroom Classic and DxO and as a computer graphics “artist” (I almost want to say “technician”) I want to work with images in Photoshop, web editors, tools I write to create images, etc.

Shouldn’t I be able to make music with GarageBand and then listen to it in iTunes and then write a program that plays it through my smart speakers at sunrise to wake me up?

Office 95 revolved around COM which meant that a Microsoft Word file was a composite file that could also contain data from other programs like PowerPoint and Excel so I could embed a small spreadsheet in a word document. (The fact that this system was documented and open was a weakness as much as a strength because you never knew if the recipient of a file had all the applications to open it)

Currently Office uses a documented XML and ZIP based file format. It is easy-peasey to load data in Excel format into pandas to do data analysis (less error prone than CSV even.). It’s not hard to write a program in PHP or Java that makes an Excel sheet complete with formulas for somebody to fill in then have them upload it back to a web site and suck the data out.

Locked in data is one reason why the cloud and mobile age feels like a step backwards than forwards, never mind the possibility of losing your data because you couldn’t pay the bill or your vendor got bought by Google, etc.

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#384
post #367

Earlier quoted context omitted.

> I have virtually no ads on my Apple devices. Virtually . It's great when you log into iCloud and only have to deal with the App Store's "Suggested Content" and the Google suggested results in Spotlight Search and the misery of the default YouTube client running 30s midroll ads. Then you can make the little storage nag go away with a convenient $2.99/month payment addressable to Apple Inc. Oh, you wanted sideloading…

Yes, if you don’t want your software and services to be ad-funded, then you have to pay for them. I thought that was sort of obvious.

So you're coping. That's fine, just make sure you never use Android and discover how green the grass is on the other side. AOSP has fewer ads than iOS in it, and that's just an unfortunate fact.

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#385
post #297

A friend doing homework for a university assignment, circa Leopard or Snow Leopard, noticed that Calculator produced negative values when raising a negative number to an even power. The bug turned out to be in CFNumber, in Core Foundation. CFNumber does a lot of fiddly stuff at the bit level for performance, and one of their optimizations for exponentiation was incorrect. Somehow it was never found by tests or due to…

In Excel (and some other languages), unary minus has higher precedence than exponentiation, so that -x^2 = (-x)^2 = x^2. That can bite you (for example when calculating the normal probability density function “manually”).

In the UK at least if you don’t put the parentheses you will get flagged for a ‘bidmas error’ because indexes (exponentiation) have higher precedence than multiplication

Source: have made this mistake. Have been flagged.

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#386

Comically, I didn't even notice those wobbly numbers in the first screenshot, I was too busy noticing: - "Unicode" button label way off center - The 8/10/16 selector being off center in its own position - The indicators for bits 31 and 63 are not aligned with each other - x and + not being horizontally aligned (I believe this is an icon-font issue, seen on HN before so knew to look for it)

It looks like a summer intern’s project made in Visual Basic.

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#387
post #372
post #303

Earlier quoted context omitted.

Interesting, I started playing with spotlight and typing in (-20)^21 returns " = 0", which is obviously not correct. And typing in "(-22)^21" gives "-71100888972574851072", but wolfram alpha insists it should be "-15519448971100888972574851072". Looks like there are still bugs here.

Hah, macOS Sonoma 14.6.1 on M1 = 0 iOS 17.6.1 = -0 WTF.

Yes, it looks like spotlight math is broken on both.

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#389
post #302

Earlier quoted context omitted.

Maybe it will be fixed in Apple Calculator for iPad (soon to be released 2024) in iOS 18?!?

Of course not. Apple would not release a calculator for iPad that doesn't honor the bigger form factor and Apple's uncompromising design and user friendliness, nothing that doesn't make users go "Wow!" in slack-jawed amazement! It needs to work on it for at least five more years, meanwhile you can buy one of the many inferior iPad calculator apps that are not hindered by Apple's vision of greatness.

This is about the beta iPad calculator in iPadOS 18 from a month ago:

https://www.macrumors.com/guide/ios-18-calculator-app/

Re: Is my vision that bad? No, it's just a bug in Apple's Calculator

#390

Earlier quoted context omitted.

Acording to https://en.wikipedia.org/wiki/Windows_Calculator#Windows_9x_... , the Win98 version of calc.exe got an arbitrary-precision math library for basic math functionality to fix this sort of problem. also mentioned by Raymond Chen, https://devblogs.microsoft.com/oldnewthing/20040525-00/?p=39... : "Today, Calc’s internal computations are done with infinite precision for basic operations (addition, subtraction, m…

Can this be true? Does the calculator have any irrational number? What if I enter PI * 2? Infinite precision, wouldn't that mean running until your RAM is full? Or would it simply work, because no irrational number is calculated to infinite (as many as the device could) digits and is therefore finite and can easily be multiplied? Or another even simpler case: If division is infinite precision, and I enter (2/3) * X,…

For 2/3, it does indeed work with fractions internally. For pi, it uses some floating point representation I think, you get about 50 digits.
Post reply on HN