1. 4

I find those articles interesting as they allow to make more informed decisions. The in-depth analisys is intriguing, a useful addition would have been (perhaps in a different document) to talk about on-boarding, e. g. how to start this process for yourself. But all-in-all it sounds like a great project.

1. 4

I personally find it hard to keep at something for almost an entire month, especially when the planning for Christmas is getting more intense. Good luck to you with pulling through!

1. 1

I know what you mean, but with the latest lockdown where I live, there’s not much to do or plan for this year anyways.

1. 2

The font choice is interesting and probably the most controversial of the points made in the article. I’m glad that this is discussed more prominently recently, but there is no satisfactory answer to my taste.

It appears that if you want a font stack that doesn’t load remote fonts (which I wholeheartedly agree with) the best resource is a blog post from 2010. After that it appears that web fonts took off and hardly any serious effort has been devoted to that. I find that unfortunate.

In the end it might the most reasonable approach to just leave the font settings untouched, as suggested in the article.

1. 6

I can kinda buy into not recommending remote fonts for bloat and/or privacy reasons, but recommending sans-serif fonts just because they may not look nice on lower resolutions feels like advice from 1999. Remember that the default typeface for Netscape Navigator was freaking Times New Roman.

1. 3

sans-serif fonts are generally easier for people with dyslexia or dyspraxia, too.

1. 1

That’s a very valid point. I’m personally for people being able to override styling to fit their needs. I’m not prepared to design my own site just for this circumstance, but if someone who faced accessibility issues raised concerns I’d be very happy to accomodate them.

Related - does dark mode help with these issues?

1. 2

Related - does dark mode help with these issues?

I don’t personally know. The British Dyslexia Association says that dark type on an off-white background is best, but they do not cite sources for this. https://www.bdadyslexia.org.uk/advice/employers/creating-a-dyslexia-friendly-workplace/dyslexia-friendly-style-guide

Edit: This paper finds that reading is faster with a warm coloured background than a cool colour. Bizarrely, they didn’t compare with white or off-white backgrounds, but it’s some empirical evidence.

1. 1

Users can set system/browser preferences to request alternate color schemes or use different default colors. I addressed both in the article.

2. 3

Remember that the default typeface for Netscape Navigator was freaking Times New Roman.

Wait, isn’t that still the default for most browsers (or at least a libre-equivalent of Times)?

1. 3

I just whipped up a test page[1] and don’t you know it, you’re right.

I see TNR as the typeface version of “PLACEHOLDER CHANGE ASAP”. People shit on Comic Sans, but at least choosing that is a deliberate choice. Leaving TNR as the default is just incredibly lazy.

1. 2

What’s so wrong with Times New Roman? I see it used quite a lot, from leaving it default in professional blogs[1] to using it explicitly as a design choice[2]. And to me it looks good.

1. 3

I might have typographical PTSD from overexposure of TMN from the mid-90s. Both sites look tolerable on HiDPI displays, but the kerning of the apostrophe in A Man’s World is a disgrace.

2. 1

The default serif font is usually Liberation Seirf, Tinos, or Times New Roman. The default sans-serif is usually the font used by the UI toolkit or whatever fontconfig has ailased sans-serif to; in other words, the system or user’s preferred font.

3. 2

the default typeface for Netscape Navigator was freaking Times New Roman.

This is a flaw, IMO. TNR isn’t optimized for displays. Most serif aren’t as readable as sans-serif fonts on low-res displays; not everyone’s using a high-res 96dpi+ screen.

4. 1

Thanks, article updated.

Diff.

1. 1

Along those lines, I’d also recommend rsvg-convert, a handy tool that I usually use to convert svg files into pdfs. It’s written in Rust and apparently part of the gnome project, if I understand correctly. In my experience it is a really fast tool for that job, which makes it nice to use if you happen to run across that use case. I haven’t tried other converting options so far.

https://wiki.gnome.org/Projects/LibRsvg

1. 5

My handwriting is absolutely the product of years of hand-written math note-taking since college. Notice the blackboard-bold and calligraphic fonts, which are also very common. Some of my own thoughts / habits:

• I can’t for the life of me write ζ or ξ. Many professors share this deficiency, making chalkboard lectures on complex analysis quite difficult to follow. Recently, I have been using the Japanese る and そ in their place, and no one even notices!

• For the purposes of communicating with others, I am overzealous with parentheses. I’d rather have them and not need them than need them but not have them. I also use whitespace quite frequently to group different parts of a longer expression.

• I never use the prime (’) symbol. Some authors use x’ (read “x prime”) to indicate a quantity which is conceptually similar to x, but I prefer to use numbers (x_0, x_1, x_2, …) or well-known letter groupings like (x,y,z) or (p,q) or (f,g) or (α, β) or (\phi, \varphi). I avoid using prime for differentiation as well, preferring the notation ∇_x f or D_x [ f ].

• Redundancy is good, and helps catch mistakes. Ideally, it should be possible to understand the gist of my proofs even if all the mathematical expressions are deleted.

• “Put a hook on the x to distinguish it from a times sign.” – not really needed

• “Put a loop on the q, to avoid confusion with 9”. – also unnecessary, since mathematicians use exclusively the digits 0, 1, 2 !

• The number two shouldn’t be written with a curl, otherwise it’s easily confused with δ or α. As long as the letter Z is crossed, there should be no ambiguity.

1. 5

Redundancy is good, and helps catch mistakes. Ideally, it should be possible to understand the gist of my proofs even if all the mathematical expressions are deleted.

Coming from a different background (implementing cryptography), I agree with this but for another reason: Readers of papers may not even have enough mathematical background to discern the notation because they come up from (possibly hobbyist) programming, rather than down from math. Therefore, having redundancy not only helps you catch mistakes, but also helps familiarize people with notation they’ve never seen before.

1. 4

Yes, absolutely! I”ve once spent two days debugging a point multiplication on an elliptic curve because x / 2 on an integer did not mean a division by two (bit shift) but rather the multiplication of the group inverse of 2. This was a really frustrating error although the solution to it proved to be quite gratifying.

2. 3

Could you post a direct link to the image? My browser is unable to render the Imgur web page.

1. 2
1. 2

thanks

1. 5

For custom written rules, I personally find that nothing beats the simplicity of redo. The build system is language agnostic in that you can execute a build step in any language and redo only tracks what needs to be rebuilt. For the example at hand, you could write the conversion rule as (in a file named default.pdf.do):

svg2pdf $2.svg$3


Then by calling redo-ifchange *.svg the entire process would be taken care of.

I would recommend using rsvg-convert -f pdf $2.svg >$3 in the example above. It is a much faster way to convert svgs to pdfs. If I remember correctly, it’s part of librsvg, an svg library written in rust.

1. 4

What implementation are you using? If I were going to use Redo, I would want to keep it in-tree because it’s not nearly as widely deployed as Make/Ninja.

1. 4

One of the nice things about redo is that there’s a single-file pure-shell implementation that only knows how to call the all build-scripts in the right order, which is great for shipping to end-users so they can compile it and move on with their lives.

Meanwhile, developers who are invested enough to build your software multiple times can install redo, which does all the incremental-build and parallelisation magic to make repeated builds efficient.

1. 1

1. 3

Not your parent commenter, but maybe they meant https://github.com/apenwarr/redo/blob/main/minimal/do

1. 2

Aha! Thanks. Currently I am using a build script by our very own @akkartik, which has worked for me so far:

#!/bin/sh -e

test "$CC" || export CC=cc export CFLAGS="$CFLAGS -O0 -g -Wall -Wextra -pedantic -fno-strict-aliasing"

# return 1 if $1 is older than _any_ of the remaining args older_than() { local target=$1
shift
if [ ! -e $target ] then echo "updating$target" >&2
return 0  # success
fi
local f
for f in $* do if [$f -nt $target ] then echo "updating$target" >&2
return 0  # success
fi
done
return 1  # failure
}

update_if_necessary() {
older_than ./bin/$1$1.c greatest.h build && {
$CC$CFLAGS $1.c -o ./bin/$1
}
return 0  # success
}

update_if_necessary mmap-demo
update_if_necessary compiling-integers
# ...

exit 0

1. 3

Appreciate any suggestions or criticism: http://karolis.koncevicius.lt/

1. 2

I said it the last time I came across your website, but I really like it! Perhaps another line with the date posted/updated below the title could work for your layout, as others have mentioned that it is missing from the blog posts.

Also, I have actually used redo for something quite similar to you, as preparing graphics for a paper/thesis recently, what an interesting coincidence!

1. 2

I really like the look of it. Very nice and clean blog post headings, though it wasn’t immediately obvious to me that the title was a link back to the index. Nicely done.

My main criticism is a big too much vertical whitespace, especially before headings. I initially thought one of the articles ended a bit abruptly, but when scrolling down it was just a large vertial whitespace before the rest of the content.

1. 1

though it wasn’t immediately obvious to me that the title was a link back to the index.

Yup that’s the main thing I cannot decide on - find a nice and minimal way to link back to the index, without having a separate “nav bar”. I considered a lot of options, from the current one, to the “X” on the corner, to the back arrow at the bottom of the page. Any suggestions are very welcome!

My main criticism is a big too much vertical whitespace, especially before headings.

That’s a good point. But when I tried to reduce the vertical space for new headers they seem to stand out less. Maybe I can reduce them just a little bit….

2. 2

Nice! My critique would be that there’s too much padding on all sides of a code block. I would cut these paddings in half or more - try viewing a code block heavy post mobile, where a larger share of the line ends up as padding than as text.

1. 1

Thanks for the comments. Padding on the code blocks are intentional - there is as much space between text and border as there is between border and code. But you are second person saying that there is too much space vertically. Will see if I can do something about it.

2. 1

I think it can help readers to at least know when a page was last updated - that way they can know instantly whether it is likely to be current information or not which matters for technical articles.

Other than that a simple “about me” page might help give some context on who you are and what you’re about.

1. 1

Yup good comments - dates can be useful. They are included in the HTML meta tags actually (both the date of creation and the date of last update). I will probably add them somewhere one day. Currently trying to be as minimalist as possible, so no tags, no dates, no RSS, and no “About me”… Want to wait and see which of these things I cannot live without.

1. 4

https://rubenvannieuwpoort.nl is my site – or maybe more a collection of articles. I tried to keep the styling minimalistic, use no javascript (math is statically rendered).

• What do you think of the layout on mobile? I like the style but feel that the text is maybe too tiny to read comfortably. I have spend some time trying to fix this but haven’t succeeded so far (it can’t be very hard, but web dev is not my niche).
• Would you prefer dynamically rendered math?

It is made with a static site generator I wrote myself (the source can be found on https://github.com/rubenvannieuwpoort/static-site-generator). Ironically, it uses node.js (and bash) and renders markdown to html.

I deliberately removed the dates from the blog posts since I tend to make large numbers of minor adjustments.

1. 4

Looks very simple, loads fast and pleasing to the eye. The equations look really cool and well rendered on my machine.

Few things:

1. It doesn’t seem responsive on the mobile. Needs the meta content viewport tag.

<meta content="width=device-width, initial-scale=1.0" name="viewport"/>

to turn it into this: https://i.ibb.co/Cbn90jC/Screen-Shot-2020-10-26-at-01-16-34.png

1. The blog post layout led me to believe that it is a PDF file because of the article being inside a box surrounded by a sea of gray. It could be a deliberate design decision, but that’s what it felt.
1. 1
1. Thank you, this is exactly what I wanted but didn’t know how to!
2. Indeed the design is based on how PDF’s are displayed. However, it’s not meant to be confusing. Maybe I’ll just get rid of the gray background.
2. 3

I really like the layout and style of the blog as it is, looks similar to the aethetic of Tufte’s work. My opinion is kind of the opposite to that of animesh in that I think the color surrounding the article works well.

The statically rendered math is amazing, I would love a solution like that for myself. Without any experience in web development I’ve not had success with it, but your blog works really well!

1. 3

This I’m struggling to read. The font is too thin, and even scaling up a lot it stays too thin for me to read. I had switch to Safari and activate Reader mode to read this. (I normally use Firefox but for some reason I can’t activate reader mode on Firefox for this website!)

1. 1

I’ve got the same issue, iPhone se 2020, with large letters (accessibility) turned on, when I zoom in, I have to scroll right and left, no word breaks. Except for that, cool minimal style and interesting content!

1. 1

Thanks for mentioning the problem and the kind words!

2. 3

Out of the ones I saw so far I like yours the most.

One question - I noticed that you do not link back to your home from within the articles. Curious if you simply didn’t find an elegant way of doing it, or is there another reason?

1. 1

Thanks for the kind words. I basically wanted the layout to match that of a printed article as close as possible. This also inspired the PDF-viewer-like look.

2. 3

You could consider putting a “published date” and a “last updated date”. I use that approach personally.

1. 2

I like the design and the content. I’d subscribe if there was an Atom/RSS feed.

1. 1

Thank you! Your words make me really enthusiastic to write something :) I might consider making a feed, but I don’t feel like I will have time for it anytime soon.

2. 2

The thin font and low contrast colour (?) make this quite hard for me to read on my phone (Firefox, android).

1. 1

I assume that this is just for the overview page, not for the articles themselves?

1. 1

Yes. I didn’t click on the articles at the time. Trying now, the text is very small on mobile. It might be that you can’t fix that while maintaining the mathematical article presentation style.

2. 2

It’s absolutely beautiful but there’s something wrong in the CSS. On my phone (Safari on iOS) it’s all “zoomed in” (I can’t see the full width of the page even) at certain zoom levels (for instance when I set 115% at the left of the address bar with the “aA” button).

Edit: I think one of the main issues comes from here:

	article {
…
width: 715px;
…
}


I would use max-width instead of width. max-width makes the width flexible for small displays.

1. 1

Thank you! I will look into this :)

2. 2

Like the simple layout, and overall clean feel. Saw someone already suggested adding the meta tag to make it responsive. Also, I feel the contrast of the articles description text on the homepage is low: light gray + light font = hard to read

1. 1

Hi all! My site’s at https://apas.gr and the blog at https://apas.gr/posts - would love to get some feedback. Thanks!

1. 2

It looks quite nice, I guess it’s heavily inspired by the tufte layout? For the posts I would suggest only showing a summary of it in order to give a better overview of the single articles. Currently it is a page that displays all your articles in its full-length.

1. 2

Thanks. Yes, indeed it’s based on Tufte layout. In fact, I built a minimalist static blog generator based on Tufte and this is a pertinent Wordpress theme. Will look into the posts layout - thanks!

1. 1

Mine is at https://secluded.site. I’d love to get some suggestions for improvements!

I recently started with Emacs and fell in love with Org mode; I’m seriously considering ditching Hugo and going much more minimal with a small, handwritten stylesheet and HTML pages generated from Org.

1. 2

On Firefox android, the pages always start in dark mode and then pop to light mode after remembering my preference (set by tapping the icon once on the home page).

Seems too low contrast in dark mode.

I also read Butterick and like the circle links and gradients, I think they’re fun though I don’t know how many people would miss them.

1. 2

I came back to this thread specially to say I love the pipe section on your site. I hope you get around to reviewing the pipes and tobacco, I’m now reading your pipe origin articles.

I see you do rss feeds per category, I like that a lot. I do it myself on my site for each and every tag, but had never seen it elsewhere.

1. 1

Hugo is what I currently generate the site with and it actually has feeds for every taxonomy, not just categories

I hope you get around to reviewing the pipes and tobacco

Once I’m finished documenting my email setup, I’ll get to work on some of those!

2. 2

Just as a heads up, you can also write your blog posts in orgmode and have hugo render it (if you weren’t aware).

1. 2

Beautiful. I like the theme and the fonts. There is one annoyance though: the system fonts are displayed momentarily, before it’s swapped out to use your fonts, but I think that’s an unsolved problem today.

1. 2

Clean, and loads quickly, but the contrats seem a bit too low for my taste. I’m not sure if this even qualifies as an oppinion, but I was expecting more on the front-page, and tried to scroll. The blinking cursor animation is actually calming!

I’m seriously considering ditching Hugo and going much more minimal with a small, handwritten stylesheet and HTML pages generated from Org.

On that topic, there is ox-hugo that can convert a org-document into a hugo-compatible site, and then render it.

1. 1

the contrats seem a bit too low for my taste

Yeah, I’ve been meaning to increase it a bit haha. Thank you for reminding me!

there is ox-hugo that can convert a org-document into a hugo-compatible site

I’m actually working on a blog post that will be exported with ox-hugo but there’s still the whole “Hugo workflow” as well as dependency on Hugo itself. Simply using Emacs and nothing else is pretty attractive.

2. 2

I enjoy your theme. The site is fast, which is really good.

I really like the way your blog posts are organized on the blog page. I like the tags at the top.

Overall the content is well formatted, with good margins.

If I had to suggest anything it would be to do something with the homepage. Use it as an opportunity to shepherd the user to relevant content, and or give them roads to walk down.

1. 2

Others have mentioned the all-but-blank homepage as well so I will very likely make some changes there.

I appreciate the suggestion and kind words!

2. 2

A few nitpicks not liking the gradients when the circle links expand; seems like “home” and “about” could be merged into one page - nothing much is going on at “home”. Also not a fan of how code blocks look in both light and dark themes.

1. 1

Also not a fan of how code blocks look in both light and dark themes.

You don’t like the colours of the highlighting or something else?

1. 1

I think they are too dark. The background is black and the text itself is quite dark within. For me it stands out too much from your otherwise light theme. They do look considerably better in darkmode, now that I took a second look at it.

2. 2

Reviewed on iPhone X.

To me, the low contrast and short line-length makes reading uncomfortable. I suggest a slightly smaller font so more words fit on a line on portrait mobile, and increasing the line-height and text contrast to compensate for the smaller size.

The link treatment was surprising, but delightful. A great touch for a personal site.

1. 1

After seeing the other post written in forum software (https://lobste.rs/s/nsdgwk/surprising_new_feature_amd_ryzen_3000) I’m wondering whether plemora could also be (ab)used for this purpose.

1. 3

I’ve stated working on a Russian translation of my compiler series. It’s going quite slowly, particularly because I’m probably at about 10 WPM in Russian, and my English articles take a long time even at 120+WPM. I’m also not particularly skilled in Russian, so I’m dreading publishing the content since I don’t know of anyone who can proofread it.

Concurrently, I’m working on an article about “how many values a boolean has”, inspired by an interview question my friend was asked. The draft is here. It’s in English for now, so at least I’m making decent progress.

1. 4

The obvious answer to your latter blog post is nine! For hardware description languages like VHDL or Verilog, there exist multiple values, as they have to take account for the actual current. IEEE1164 gives a small overview if you’re curious: https://en.wikipedia.org/wiki/IEEE_1164

1. 3

Wow, thank you! I have been taught some SystemVerilog and CMOS stuff, and recall having “high”, “low”, and “somewhere in-between” (forgive me if that sounds silly - I’m no electrical engineer!). However, I certainly did not know about all of these options. Thanks again for sharing!

1. 2

Yes, I was in the same situation, and I remember being irritated when this table was shown to us. Happy to help :)

2. 1

Hi Danila, I am in a similar situation to you with regards to wpm in Russian.

I highly recommend you look into “translit” keyboard layouts, in particular the one designed by Apple.

On macOS, it’s known as Russian (Phonetic).

For Windows, a kind person has created an adaptation: https://medium.com/@yurinnick/battle-around-russian-phonetic-keyboard-in-windows-10-c9cfbe3be707

For other platforms, probably have to hack it yourself.

Я не могу сказать что печатаю так-же быстро как по-английски, но по крайней мере я знаю где все клавиши. С привычной клавиатурой, я-бы сказал моя скорость–40-60 знаков в минуту.

1. 4

I am writing my master thesis at the moment. Submission date is in less than two weeks so this is my main focus for now. The topic deals with dimensionality reduction (read: visualization) algorithms, to be specific with neighborhood embeddings. While conducting research the focus was on a comparison between t-SNE, UMAP, and ForceAtlas2 (and to a lesser extent Laplacian Eigenmaps). In general those techniques are used for transcriptomic data quite a lot but there is hardly any intuition or theoretical underpinning behind it. So the extent of my work was doing a comparative study from an analytical standpoint as well as empirical with plenty of ablation experiments.

Now it’s mostly writing it all down – with a special focus on the mathematics behind it –, finding (even more) references, and making some additional illustrative figures.

1. 6

This sounds a bit like it’s going to take the course of USB3.0, which turned from an improved USB into an abomination that tried to do everything. While that certainly not only has downsides, I’m not sure if it’s a good development.

1. 2

Yeah reading “5G ready” on an antenna will hardly say which part of the standard will be supported. Will you be able to connect to short-range frequencies (30 - 300GHz) as well? Unlicensed spectrum (5-6 GHz)? It is certain that when buying a phone for example this kind of info will be usually missing, just like USB-C does not give exactly what is supported by the port.

On the other end, industry features (sidelinking, V2X etc) will change a lot of things, not only for corporations but also citizens and society. Good or bad remains to be seen.

1. 3

I love the data type graphs and find it amazing how expressive it is and how naturally it can be used for modelling just about anything. I am currently writing my thesis on using Neighborhood Embedding algorithms (basically an algorithm operating on the distance matrix between datapoints) for visualization and I find the whole topic fascinating.

That being said, by virtue of being so flexible, they can, generally speaking, be computationally expensive to use, as without any restrictions just creating a graph would cost O(n^2) space and time. In many application domains it is possible to exploit the structure (usually sparsity) in order to push this down, however.

All in all, I think graphs are a powerful data structure and can become more important in the future, especially if theoretical progress continues to be made.

1. 1

To me it seems that annoy is more or less an industry (research) standard. In general, it doesn’t seem that speed is the bottleneck for ANN queries, but memory requirement is excessive, at least from what I observed.

In general, I don’t know of many useful applications that require high-dimensional features, from what I have done so far (which is nowhere near extensive) it seems to work quite well to reduce the data to a lower dimensional representation via PCA (say from 10,000 to 50) before proceeding. At least this seems to be quite standard in the transcriptomic community. So I am not entirely convinced that it’s necessary to be able to accommodate this many dimensions in an ANN query.

1. 1

This was super interesting to read, although I have virtually no experience in fuzzers. It’s interesting that research considers probability distributions induced by fuzzers and the code (coverage). That’s quite a unique way to look at source code and testing. I wonder if this could be used to prove a program correct, with convergence almost everywhere or something to that extent.

1. 2

I’ve been using this option for quite some time. the initial setup is really no more complicated than setting up another ssh host (and optionally generating another key for it). The great thing is that this also allows you to keep your private keys on the computer where you initiate the connection instead of having to create them on the ProxiJump hosts (which may not always be possible).

1. 2

Thanks for that, I have only dabbled in pytest so far, but was pleasantly surprised by the basic features already, so that was an interesting read.

1. 1

That is lovely, I wish I saw it sooner. The details to the approach by the Economist is quite comprehensive!