blog/content/links/20200604.md

+++
title = "Commented Links for 2020-06-03"
date = 2020-06-03

[taxonomies]
tags = ["links", "rust", "scotty", "python", "pyramid", "cornice", "datetime",
"mercurial", "async requests", "httpx", "advertising", "google", "python 3"]
+++

Scotty, Pyramid and Cornice, Problems with `datetime.now()`, Mercurial and
Python 3, Fast Python, Async Python Request Library, Online Advertising,
Google and Web.

<!-- more -->

## [scotty: Transports you to any directory you have visited before](https://github.com/wdullaer/scotty)

I have been using Scotty for some time and I really like it.

What Scotty does is keep a list of directories you accessed recently and,
after a while, you can simply use `s <directory>` to get straight into it. Not
only that, but Scotty uses fuzzy searching to find the directory that closely
matches whatever you typed.

## [How to write a Python web API with Pyramid and Cornice](https://opensource.com/article/20/1/python-web-api-pyramid-cornice)

Pyramid is the "less known" Python web framework around. But this post shows
how simple is to build a (simple) API with it.

## [Stop Using datetime.now!](https://hakibenita.com/python-dependency-injection)

Slightly clickbait title, but great content. Not only pointing out the
problems when using `datetime.now()` (no, it's not the function itself that
has a problem, it is the way we use that may cause problems), but going into
lenghts explaining dependency injection in Python.

## [Mercurial's Journey to and Reflections on Python 3](https://gregoryszorc.com/blog/2020/01/13/mercurial%27s-journey-to-and-reflections-on-python-3/)

Ding dong, the witch is dead, and so is Python 2.

But there is still a large base of Python 2 projects that need to be converted
to Python 3, and Mercurial did this, and here's the experience of a
maintained on doing it so.

I do understand that converting Python 2 to 3 is not a simple task, but there
are a few misconceptions in the post. For example, "the approach of assuming
the world is Unicode is flat out wrong and has significant implications for
systems level applications". The word _is_ unicode. Go read the Portuguese
version of this blog to have some idea. Go read any Chinese/Japonese blog to
see how it looks. Do you really think those people do not use system level
applications anywhere? Also, what do you think are mostly used: User level
applications or system level?

Not saying the conversion is perfect -- in a way, unicode is simply a way of
dealing with the underlying bytes -- but ranting that this change made _your_
specific way to think harder is not an excuse to not understand where the
whole ecosystem was moving -- and it also doesn't mean your specific thing is
not welcome, but you have to understand you're in the minority case here.

(Also, it's no surprise to me that claiming "world is unicode is wrong" is
coming from someone living in an ASCII country.)

## [Making Python Programs Blazingly Fast](https://martinheinz.dev/blog/13)

A set of tips on how to make Python applications faster. Some of those may
seem weird in the first glance (make a variable in the function point to the
a variable in the same class?) but the post also explains _why_ this may
improve the general performance.

## [Httpx: A next-generation HTTP client for Python](https://www.python-httpx.org/)

[Requests](https://2.python-requests.org/en/master/) shown to everyone how
APIs should work, but it kinda got stuck in time, without async support (and a few
other glitches in the project management). Now there is Httpx, which is,
again, another HTTP request library, but this time with async support and,
basically, the same interface.

## [No-judgment digital definitions: Online advertising strategies](https://blog.mozilla.org/firefox/online-advertising-strategies/)

Have you ever wondered why after you search something -- say, "gamer chair" --
suddenly you start getting a lot of promotions and ads for chairs in your
social networks?

Nothing happens by simple chance, and that's the effect of all the trackers
that someone puts on a page. But how they identify who is who is the real
question.

This post by Mozilla may seem a little bit basic, but shows pretty damn well
how those things are done.

## [Google Is Not God of The Web](https://bilge.world/google-page-experience)

Another clickbait title but, again, good content.

A point that caught my attention was "Google has the right to dictate 'Best
Practices', although I think the topic is quite the opposite, based on its
content.

Imagine that Google start giving points of "user experience" to pages that use
the Material design. Pages and pages that look like Android apps. But if you
use your own layout, your own colors, you lose points. What now?

This is the greatest problem on people depending on Google, and all the
relationship of one of their groups working on web standards, a group working
on a browser and a group to take advantage of whatever the previous two did.
The less dependent on Google you become, the more you use alternative search
engines (DuckDuckGo, Searx and even Bing), the more you use other email
providers (ProtonMail), the more you use browsers that are not Chrome
(Firefox, Safari or anything based on WebKit), the lesser the chance of the
internet becoming the thing of a single company.

---

This post was built with the help of

* [HN Tooter](https://mastodon.social/@hntooter)