bfm

The OP details how poor software engineering practices brought down a 1.4B market marker with 1400 employees in 2012.

Some of the issues mentioned include:

  - Keeping synthetic test data generation as part of a production build.
  - Keeping dead code for years.
  - Re-purposing a feature flag.
  - Refactoring without regression tests.
  - Manual deployments without peer reviews. They forgot to update one of their servers with the new code.
  - Automated alerts sent via email were ignored.
  - Rolled back to a version of the code running on the server they forgot to update, making things worse.
  - Rushing out a release without proper software engineering hygiene.

The article suggests improvements that could have prevented the chain of events.

For those here who are in HFT circles, have things improved after the Knight Capital Group debacle?

edit: formatting

It looks like the IMPORT_CACHE can be used for vendoring.

A quick example:

    #!/usr/bin/env bash

    export IMPORT_CACHE="vendor/import" 

    source ./bin/import

    import "tootallnate/hello"
    import assert@2.1.3

    hello
    assert_equal foo bar

generates the following dir structure:

    vendor/import/
    ├── data
    │   ├── 0a1c5188c768b3b150f1a8a104bb71a3fa160aad
    │   └── a4f0a20c03de97c490413f0c58f480f2ad17ba27
    ├── links
    │   └── https
    │       └── import.sh
    │           ├── assert@2.1.3 -> ../../../data/0a1c5188c768b3b150f1a8a104bb71a3fa160aad
    │           └── tootallnate
    │               └── hello -> ../../../../data/a4f0a20c03de97c490413f0c58f480f2ad17ba27
    └── locations
        └── https
            └── import.sh
                ├── assert@2.1.3
                └── tootallnate
                    └── hello

After years of using Terraform and Puppet, I seriously consider using plain bash to handle infrastructure as code on a future project, which deploys to multiple cloud and non-cloud vendors.

There is some beauty to creating a standalone idempotent and well-tested bash scripts that are simple to share and version using something like import.sh

Given the maturity of tools such as shellcheck, simple composable tools such as age (for encryption), machma (parallel execution), Make for the dependency graph and the assistance provided by Github Copilot to developers intimidated by bash syntax, what would be the drawbacks of going with this approach?

What other simple tools like import.sh would you recommend for building IaaC pipelines?

This!

After working with node.js for more than eight years I decided to give deno a try (https://github.com/bermi/genetic) and I’ve been gratefully surprised by the development experience.

I’m just missing a simple way to include deno code on my existing node.js and browser projects.

If you have not been following the work José and Sean have been doing lately, I recommend you watching José’s talk introducing Elixir Nx https://youtu.be/fPKMmJpAGWc

A quote by Avi Bryant regarding continuations in Drabble DB "... in Dabble we essentially don't use continuations although we do use Seaside"

You can read the whole interview on https://www.infoq.com/interviews/bryant-smalltalk-dabbledb/

Good luck with the launch!

I've been DIY'ing exactly this full time over the last year. Rolling out my own system using Polygon and IEXCloud for market data, DolphinDB and Shakti for time series processing, and Pandas/Dask for verification and parameter optimization.

I totally underestimated the magnitude of the problem, so I'm really looking forward to trying Enombic.

Given my frustration with the DIY approach, I started evaluating some prosumer tools a couple of weeks ago. One of them uses Morningstar data, and the other uses FactSet data. They are in the $100-$200/month range. I had some decent success building statistically sound models to the point where I'm considering dumping my DYI toolset and pay for a ready-made solution.

One of the alternatives has an active forum and community that even sells subscriptions to the user models. I bet you can tap this for your own good. I won't share their names here to avoid crashing your launch, but I'm happy to chat over email if you want to learn more.

Some of the questions that I'd ask myself before putting my savings on autopilot on a system like Enombic are:

1. Where is fundamental and price data sourced from? Is it free from survivorship bias? Is it point in time? 2. Can users backtest/simulate their ideas? How do you help users from curve fitting their models?

Anyhow, I'm on the waiting list now and would love to evaluate Enombic and give you my feedback.

You're probably right. After reading through https://twitter.com/TwitterSupport/ it looks like Twitter has been disabling some features. And according to https://twitter.com/louanben/status/1283518716118958080/phot... the @TwitterSupport account was also affected. I doubt they use a system like HootSuite for that account.

And their status page shows their integration with Twitter is having issues now https://web.archive.org/web/20200716000356/https://status.ho...

Looks like Hootsuite Twitter integration has been having issues for 50 mins now https://status.hootsuite.com/post/623750375373160449/twitter....

You also have Backbone events extracted from Backbone on https://github.com/bermi/eventify

From the official announcement https://groups.google.com/forum/m/#!topic/nodejs/fqBTNLwhEAI

    The main advantages/differences over Connect/Express at this   point are:

    - Patterned after WSGI/Rack. This implies some significant differences in the API.
    - Built on node 0.4; takes full advantage of streams.
    - Built-in support for streaming multipart parsing/file uploading (uses node-formidable's parser).
    - Built-in support for gzip encoding, URL rewriting, and URL mapping.
    - RFC-compliant support for content negotiation using the Accept, Accept-Language, Accept-Encoding, and Accept-Charset HTTP headers.
    - A specification with middleware to enforce it.

I'll add favorite colors support in the future.

What browser/OS where you using when you had the download issue?

Thanks for the suggestion. I just changed it.

Event driven development was instrumental. Back in 2007 I coded a similar project in PHP and had to take the app offline because it was painfully slow for end users.

I'm sure you can do something similar in any language by using something like RabbitMQ.

Thank you!

I learnt Node.js mostly by looking at projects on github, reading Node.js API http://nodejs.org/docs/v0.4.2/api/ and following these two great resources http://howtonode.org/ and http://blog.nodejitsu.com

Thank you. The app is hosted on http://no.de Its a Solaris server with just 128RAM and 5GB HD

Thanks for the suggestions this is something it will certainly work on.

Thank you!

This project is just for learning and fun. No ideas yet for monetizing it.

I'm sure this can be done using Canvas. I might give it a try as a learning exercise in the future. Thanks for the suggestion :)

Didn't know MapRejuice, looks promising.

Regarding no.de, I have the same issues you describe. I'm looking forward moving to a larger SmartMachine where I can deploy using capistrano.

Never used Lightroom, but the presets on http://goo.gl/wQ0PV seem to be formatted in json.

Seems like this presets can be created easily from a list of colors.

Interesting application.

Thank you!

Thank you

Thank you Michael for the feedback.

So far I've just tried it on FF, Chrome and Safari on Mac. Cross browsing is on my list.

Yes image processing happens in a child process.

Right now I'm using GraphicsMagic resource limiting flag -limit for preventing it from taking down all the resources.

Some background on the application http://www.bermi.org/announcing-colorapi-search-for-colors-i...

I wonder who leaked my password http://www.apathysketchpad.com/images/uploads/mushport.jpg