explosion

While working on a commercial project, you typically have to write some code that isn't related directly to the core business. This code can be open sourced as a library.

That's great to hear. Thanks.

In the shared runner settings, I see this:

"GitLab Runners do not offer secure isolation between projects that they do builds for. You are TRUSTING all GitLab users who can push code to project A, B or C to run shell scripts on the machine hosting runner X."

Seems like a very strong reason to use one's own paid DigitalOcean instances for runners instead of using the free shared runners, at least for commercial projects. I was wondering if anyone from GitLab could expand further on this?

I remember taking AP Computer Science in high school and learning Pascal. Anyone else remember doing this? The AP Comp Sci exams were in Pascal until 1998.

Is there any reason why one can't be both a writer and engineer simultaneously?

I see, thanks. Makes sense.

To be fair, it was misleading that the WSJ article said "Google is changing its search algorithm Tuesday to favor sites that look good on smartphones."

By hand meaning "pip install -r {file}" or more generally pip with any arguments typed directly at the command line.

Automated meaning pip with any arguments run by another script.

What would make more sense than blindly prioritizing mobile sites is this:

1) Desktop-friendly websites appear higher in the rankings for searches made from desktop computers

2) Mobile-friendly websites appear higher in the rankings for searches made from mobile devices

I wish there was a way to tell which pip installs were done by humans vs. by automated tools. It would be interesting to track what people are doing by hand vs. what machines are doing. I suppose that may be impossible to implement, though.

These graphs are very insightful. Thanks for offering us some visibility into the Python package download picture.

This survey was taken on CodeProject.com, a site which seems biased toward C#, C++, .NET, Java, and mobile devs.

Perhaps it's just that sample of developers who skew toward INTJ, ENTJ and INTP.

I suspect that the results would be different in a survey of developers from a more open source-focused community, e.g. if GitHub conducted that survey and a lot of idealistic open source developer types responded.

I agree, the resulting images are beautiful. For some reason, it's surprisingly mesmerizing to drag the points around and see the colors change.

I like that tool very much, but it would be nice to see the generated CSS styles displayed separately from the HTML.

By this, I mean generating the following individually: 1) the HTML with a class applied to a sample element to be centered 2) the separate CSS for that particular class selector

This is just a little nitpick based on my preference for avoiding inline styles; overall, nice work!

That phrase is new to me. ikeboy, thanks for the tip.

I'd be interested to hear more stories like this about seeding sites successfully without appearing fake. I've seen a lot of startups seed their sites in ways that just look too obvious.

wcaleb, this is great, thanks for both posts. It's also very inspiring to see your open notebook workflow and experiments.

Another huge reason to use plain text that he didn't mention is version control. With plain text, you can check the files into a git repo and diff any two commits. You then get all the advantages of working with version-controlled code, e.g. commenting on diffs.

Word and Google Docs do save revision history, but they slow to a crawl when you have a document with a lot of pages, to the point of being unworkable.

Fascinating. Curious what it would take to register, I found the Pyongyang Marathon signup form here: http://www.pyongyangmarathon.com

Apparently they accept amateur runners and you can do a 10k, half marathon, or full marathon.

It's good that they're planning for their future storage needs to increase drastically. I suspect that 1.2 petabytes will seem like a gross underestimate in 2025, though :)

I like the author's concept of PACELC, though it seems a bit implementation-specific.

Going from working for a company to freelancing has definitely increased my number of daily work hours.

However, in between contracts I have the ability to take long breaks to work on my own projects or travel. Overall, the benefits have outweighed the risks for me, but then again, I'm just one data point.

I'm having trouble understanding how one would demonstrate traction at the idea stage.

Would it be a combination of 1) demonstrating past successes gaining traction and 2) having a catchy prototype? Or are there other things?

Or perhaps they're just trying to diversify their portfolio.

Perhaps similar to how while Google acquires all sorts of interesting startups, it doesn't mean they're shifting away from search/ads.

I actually think that the article's example using Dali's The Sacrament of the Last Supper sort of disproves itself.

Part of why the painting is so elegant is because of its balanced composition: the arms along the top of the golden spiral, the table edge along the top of the inner spiral, etc.

I noticed how even though your book doesn't exist yet, the graphic that you put up on Kickstarter looks very polished and enticing. But I've also seen startups dig holes for themselves by over-obsessing about graphic design.

On a startup's first MVP iteration, how little graphic design effort can you get away with? Does it depend on the product, perhaps? Should all iterations be visually polished?

I'd like to see more startups and companies pledge to do this as well.

Having open-sourced a number of components and then seen them get better from community contributions with little effort, this is a good thing to do.

Nice work, and thanks for open-sourcing this!

There seems to be a lot of under-explored potential for building useful, pluggable D3 widgets along these lines. I'd love to see more interactive widgets like this.