mau

the article complains about ux, not capabilities. In this case it is complaining about the defaults.

I think humans are still accountable for the code generated by agents.

You are free to switch language but you still need to understand it.

German-style strings is a way to store array of strings for columnar dbs. The idea is to have an array of metadata. Metadata has a fixed size (16 bytes) The metadata includes the string length and either a pair of pointer + string prefix or the full string for short strings. For some operations the string prefix is enough in many cases avoiding the indirection.

This is different from Pascal strings.

https://archive.is/2025.12.29-103247/https://m.independent.i...

Your coworkers and QA will thank you for learning LINQ and ditching the imperative methods that plague your Python brain.

This is a very unfortunate joke: Python has list (and generator) comprehension expression for a long time (2.3?) which are similar to LINQ. At some point in the history many languages stole useful expressions from other paradigms.

Let’s joke on BASIC, it always works.

Arrow supports this string format: https://arrow.apache.org/docs/format/Columnar.html#variable-...

From the article:

As luck would have it, the Arrow spec was also finally making progress with adding the long anticipated German Style string types to the specification. Which, spoiler alert, is the type we implemented.

Congratulations to the team, Pydantic is an amazing library.

If you find JSON serialization/deserialization a bottleneck, another interesting library (with much less features) for Python is msgspec: https://github.com/jcrist/msgspec

I think the GH app is a workaround for this:

Deploy keys only grant access to a single repository. More complex projects may have many repositories to pull to the same server.

I don't know what were the issues Yelp was facing, I have upgraded several times my past Kafka clusters and really never experienced any issues. Normally the upgrade instructions are documented (e.g. https://kafka.apache.org/31/documentation.html#upgrade) and the regular rolling upgrade comes with no downtimes.

Besides this, operating Kafka never required much effort a part when we needed to re-balance partitions across brokers. Earlier versions of Kafka required to handle it with some external tools to avoid network congestions, but I think this is part of the past now.

On the other hand, Kafka still needs to be used carefully, especially you need to plan topics/partitions/replications/retention but that really depends by the application needs.

You can write code that runs in both interpreters. Same syntax can have different behavior.

I guess nowadays, as the py2 compatibility is a thing of the past, most programs would crash when using that interpreter version.

One thing that is underestimated is keep the tools version in sync between your app dev dependencies and pre-commit. This also includes plugins for specific tools (for instance flake8). A solution would be to define the hooks in pre-commit to run the tools inside your venv.

About typings: I agree the eco-system is not mature enough, especially for some frameworks such as Django, but the effort is still valuable and in many cases the static analysis provided by mypy is more useful than not using it at all. So I would suggest to try do your best to make it work.

We are currently using Kafka Manager (https://github.com/yahoo/kafka-manager). It seems Kafka HQ have the same features (it is not clear from the doc if you can actually manage the partition arrangement or only view it) plus the possibility to view the actual messages that are streaming in Kafka that might be very useful.

https://developers.facebook.com/docs/audience-network

FB Audience Network is already in Beta for Mobile Web [1], they might investigate to expand this business for Desktop Web too.

[1] https://developers.facebook.com/docs/audience-network/mobile...

Much of the effort for the next release of Druid (0.9) has been put into docs. Even with that you still need to handle a complex cluster made of several services: zookeeper, druid overlord, druid middlemanager, druid broker, etc... still it sounds more complicate than it actually is. What was the alternative solution adopted by your company?

Here is a profiler developed by dropbox using the same technique described in the article:

- https://blogs.dropbox.com/tech/2012/07/plop-low-overhead-pro...

- https://github.com/bdarnell/plop

Many people here are comparing this feature of Postgres with MongoDB or other document-oriented NoSQL dbs. I think the comparison is just wrong and unfair.

Even if I'm amazed by the performance of Postgres for this particular task (considering also that is a relatively new feature), I don't think performance is the reason why people are using NoSQL dbs. The problem that NoSQL dbs are helping with is scaling. I don't see this as a priority for an RDBMS such as Postgres. Take for instance MongoDB (just because it was named by many of the other comments here, but I guess the same apply to Couch or others): it's relatively simple to deploy a cluster with automatic sharding, replica, failover, etc.. because these are all builtin features.

A framework is not only a good way to start. It's also a good way to uniform code, make things easier for new comers (larger projects usually have lots of new people that eventually write code), solve multiple common issues you'd fight by your own (e.g. browsers compatibility), community support and documentation. If you write your own you need to document well all things and things very careful how to arrange css classes to make code reusable and flexible.

Can you please add some sources for the claim about Feynman? A quick search on the web "feynman abuse" didn't gave me relevant results.

I had some "memory leak" issues a few months ago with the api hosts of the service I'm working on. As the code changes made during the days before didn't seem to have affected anything that could be related to this leak I realized it was something about the requests we were receiving but I wasn't able to replicate the same issues in my dev host nor in staging. Eventually I dumped the memory of a running process using Pyrasite and after hacking a little bit with meliae (https://launchpad.net/meliae) I was able to find the issue and solve it.

Very useful.

By the way is worth noticing that you won't probably be able to attach the interactive console to a running web server as the output is usually handled by the supervisor process, at least I wasn't able to do that in my first tentative and the memory dump was good enough for me.

Check this sample to dump the memory out of a running process: http://pyrasite.readthedocs.org/en/latest/Payloads.html#dump...

If you cut and paste probably the misaligned indentation is the least problem. Cut and paste should be avoided if it's not done during a refactoring (so actual moving of code for better design).

Beginners usually think Python strong indentation is a weakness of the language. I actually find C++ freedom being more error prone:

    if (a < b);
        a = b

Don't C compilers warn about code that is never executed?

I'm actually pretty impressed how this bug was not caught before. The compiler warning is just one thing that came to my mind but test coverage should have been the strongest hint. A linter could have also make the thing easier for a code review. Probably these parts of the code should have stricter coding guideline.

you can find something that do the 1st job here: https://github.com/danielgtaylor/qtfaststart

"Some versions of Chrome feature Safe Browsing technology that can identify potentially harmful sites and executable file downloads not already known by Google. Information regarding a potentially harmful site or executable file download (including the full URL of the site or executable file download) may be sent to Google to help determine whether the site or download is harmful. Google does not collect any account information or other personally identifying information as part of this contact, but does receive standard log information, including an IP address, URL visited and one or more cookies."

It's a bit confusing for me, it looks like they are collecting visited urls as they are not considered personal identifying information.

I've just learned that with the spellcheck feature they are basically receiving all the text I write on the Internet to their servers.

By the way are we really discussing about the privacy policy of a company that sent its data to NSA? I'm not condemning them, still we should be aware of this and never forget.

For the time I was working there I'm quite sure most of the tools were internally made by MS: - they had their own map reduce - they had their own service deployment and management system - their own db and nosql - etc...

about the search and relevance they are updated quite fast, sometimes with big code replacement. This means that even if powerset tech was used as a start (and I do think it was not, it was possibly adapted and integrated -- i believe the base code was msn/live) it is now changed in most of its pieces.

> The Winapi (aka Win32) is the only commonly used API that regularly requires something other than UTF-8 (the Windows Unicode APIs use UTF-16 – not UCS-2 as indicated in the Spolsky article).

Why has Microsoft still not fixed this?

Is there a way they can fix it without breaking backward compatibility?

tldr: the ML algorithms look hard reading the papers, while the code looks simpler and shorter, also you can get pretty decent results in a few lines of R/Python/Ruby so ML is not that complex.

I disagree in so many ways: 1. complex algorithms are usually very short in practice (e.g. dijkstra's shortest path or edit distance are the firsts that come to mind) 2. ML is not just applying ML algorithms: you have to evaluate your results, experiment with features, visualize data, think about what you can exploit and discover patterns that can improve your models. 3. If you know the properties of the algorithms you are using then you can have some insights that might help you on improving your results drastically. It's very easy to apply the right algorithms with the wrong normalizations and still get decent results in some tests.

What a coincidence. I just had a conversation with a friend about how FB failed with its Messenger to get the users of Whatsapp & other similar apps. Apparently they know this and are trying to improve it. Same discussion applies to Skype that was never able to make a fast user experience on mobile: every update of Skype on my iPhone seem to slow down the app and make it more crash friendly...

Related to this, about Google: http://bits.blogs.nytimes.com/2013/09/19/google-is-exploring...

Just to add some extra info: Wikipedia states that Connect Four was solved mathematically in 1988: http://en.wikipedia.org/wiki/Connect_Four

Few months ago I wrote some sample code to show how to implement a simple min-max based algorithm to play zero-sum game. It is a very basic c4 player but it might be helpful for someone who wants to learn this stuff. The code is on github: https://github.com/duilio/c4

How do you pre-render the pages for the client? Do you just send a static HTML and let the js populate the page or there is a better way from python to not duplicate the template processor?