I run Gogs on a Digital Ocean VPS. $5/month (plus $12/yr for the domain), dead easy, and stupid simple.

Requests even has a feature to automatically parse the JSON.

I used Nutch to map a corporate intranet and feed the data into Elasticsearch. Nutch, to me, seems like a better tool when you don't have specific sites targetted, though it can work for that.

Everything apart from the title was in the content tag as one long string.

It's been a year since I used it, but I don't remember this being a problem. Like I said, I fed the data right into ES. There was already a plugin (either for Nutch or ES) for parsing that data properly into the index.

I used Nutch for a project last year. There is a little learning curve and it requires some configuration, but as you said it's mature. It works quite well for webcrawling and can be setup for automatically pulling metadata and full text out of crawled documents and can easily dump the crawl results into an index like Elasticsearch.

That being said, it sounds like you're targeting a single website looking for specific information, which seems like more of a scraping task than a crawling task. I'd look at Scrapy.

Every time you use encryption, you're protecting someone who needs to use it to stay alive.

Best line from the article.

I have one I've used as:

1. a main desktop
2. a personal journal
3. a wifi router
4. a target to learn how to do cross-compilation
5. a PirateBox
6. an IRC bot

It's great to have a cheap-to-buy, cheap-to-run computer around for doing random stuff. I also want to collect a few RPis to prototype an idea I have for a distributed system, but I'm not giving that one away yet.

this one people literally think is the "C Bible".

In a way it is, isn't it? It was written by the people who invented the language. The first edition was the first published standard for the language. The second edition was close to a standard. (Though ANSI C was a thing by then.)

Sure there are other good books about the language. There are other, better books for learning it that include the decades of improvements in the language and ecosystem since then. But, K&R was first, it was the best of its time, and it remains a great example of technical writing, even if it is no longer the 'best' book about C, objectively speaking.

Is this supposed to fully replace JS, with access to the DOM and all that? I would love to not have to use JS when coding website interaction. I didn't think that's what NaCl was intended for, but that's how the article reads.

How long will it take for Apple and co to add proprietary APIs to this, breaking the Web (again!).

From the promoted comment in TFA:

This is exactly how the system is supposed to work: individual teams try to advance the state of the art, and eventually all those lessons learned are incorporated into a new and better system. See e.g. SPDY -> HTTP2. WebAssembly draws on both the past work and the experience of all those involved, and wouldn't be what it is without them.

As long as there is robust browser competition between 3+ major vendors, those individual browser mutations allow the whole browser population to evolve.

Not really. The bottle docs are good and the framework is simple enough that I didn't need to look for any other walkthroughs.

All I did on my last one was set up the interface points with bottle, as in the docs. For each point, get data from the request, decode it with json, prep the sql statement, run the query/update with sqlite3, serialize the result with json, and return the result with bottle. Later I added user/password authentication with passlib for salting/hashing and random for generating tokens.

For all of this, I just looked at docs of each component then glued the pieces together.

For serious number crunching Fortran (yes people still use it)

Absolutely! The whole scientific Python stack is built on top of Fortran libraries BLAS and LAPACK. I'm lucky enough to never have had to do any custom work at that level, though.

I've never liked Oracle. There's a reason LibreOffice and MariaDB forked when Oracle bought Sun. They've been pretty hostile to the open source communities they acquired (OpenOffice.org, MySQL, VirtualBox, Java).

But when it comes to Android, Google is big enough to stand up to Oracle and keep the individual Android devs safe.

1. Use Python3 if you don't already. It's the future and it's better. And, if your favorite module is one of the few that only supports Python 2.x, it shouldn't be your favorite anymore.

2. Use IPython instead of the default REPL if you can. It has tab completion and remembers what you typed in previous sessions. It's nice.

Exactly, using bottle, json, and sqlite, I can have a CRUD microservice set up in about 5 minutes. Then 2 minutes later, I can have a script consuming that service with requests. I don't think I like python for very large scale projects (though it gets used for them), but there's no quicker language for me to prototype anything.

Relevant XKCD

Like best to use? Python. It's quick and easy to turn my thoughts into code. Also, the Scipy stack is superb for my data work.

Have to use most? Java. It's mature. It's capable. It's everywhere. It's not C++.

Most want an excuse to use? Golang. Faster than Python. Still seems easy to write. Cool concurrency primitives.

This article was last updated two weeks after Bill Clinton beat Bob Dole, the Monica Lewinsky scandal was still 1 year an a half away, Ross and Rachel were still together the first time, and Python was still on the 1.x series.

I'm not sure this article is relevant.