6 ms·

National (JavaScript) Security Agency

48 points by thewavelength 10y ago

maze-le 10y ago
> assets/js/main.js I smell rails
- robin_reala 10y ago
  Default Rails would be assets/javascripts/application.js
  - _asummers 10y ago
    JavascriptS? Plural? Interesting.
    - clentaminator 10y ago
      Don't forget CSSes.
    - paulojreis 10y ago
      I also got intrigued when a fellow colleague used that nomenclature (folder called "javascripts", as well as "stylesheets", inside an "assets" folder).
- tcfunk 10y ago
  Is smelling rails a bad thing? (Genuine question)
  - maze-le 10y ago
    Nope, I don't think so. I am a rails developer myself. It might have sounded a bit condescending...
DCRichards 10y ago
This isn't as secret as you might think... http://imgur.com/VYZbZKF http://imgur.com/VYZbZKF
mpeg 10y ago
// start comment block That's some properly documented code ! // end comment block
- dagwaging 10y ago
  my personal favorite: // form setup try { formSetup(); } catch(e) {}
- 0xfeba 10y ago
  Heh. // title update try { titleUpdate(); } catch(e) {} // titleUpdate() //update title to include " - NSA.gov" on all subpages function titleUpdate() { ... }
cookiemonsta 10y ago
What would be the reason for this: // analytics disable if url contains underscore try { if(window.location.href.indexOf('_') > 0) {window['ga-disable-UA-20753548-1'] = true;} } catch(e) {} });
- robotnoises 10y ago
  My guess is that there is a test scenario in which they don't want google analytics to load?
- mpeg 10y ago
  It seems like they store files (presumably both internal and public) under https://www.nsa.gov/ia/_files/ https://www.nsa.gov/ia/_files/ so I guess it's the way they represent their non-indexable urls.
- gizmo 10y ago
  The if condition is wrong. This is why functions like indexOf should return a tuple of a boolean and an offset instead of encoding "not found" by returning -1.
  - ChristianBundy 10y ago
    The if condition should use >= rather than >, but it otherwise makes sense.
    - dahart 10y ago
      I'm not arguing in favor of writing it that way, but I might argue it's not incorrect as written. location.href can never start with an underscore. A regex on location.pathname would sure be a lot clearer to read, though.
crisisactor 10y ago
Stop advertising the NSA please
- BasDirks 10y ago
  This surely isn't advertisement. The code is terrible. Try-catch jungle, ad-hoc URL manipulation, lack of caching, writing/constructing HTML in strange places, weak comparison operators in dangerous places (string vs lunch), etc.
  - crisisactor 10y ago
    Yeah but when visiting in Lynx, it displays fine. Lynx is the yardstick I use to judge a website. There are countless .GOV sites like this made on old Windows 2000 machines with Frontpage installed, but my point is the information is still communicated. It probably doesn't matter about some spagetti JS lumped in. I assume any sensible person will have JS turned off when visiting a URL like that anyway...
    - BasDirks 10y ago
      I see. I think it's important to make websites work well for as many clients as possible, including lynx. Ideally, JavaScript should only be used for enhancement of necessary interactivity. But when a choice is made in favor of some technology, I like to see it used properly. In this case maintainability, performance and robustness could benefit from improving this code.
    - kiya0 10y ago
      :) man have you even met today's web developers? Turning off js is like asking them to work in assembly. Now that Google has gone and built everything the OS does, into the browser the mordern day web dev is programmed to think that they exist to rebuild desktop apps in the browser. Communicating information simply in plain text is the last thing on their minds.
      - BinaryIdiot 10y ago
zuck9 10y ago
Looks like it was written by a VB programmer, who's missing "on error resume next" in JS.
- rbirkby 10y ago
  Should have used RxJS https://xgrommx.github.io/rx-book/content/getting_started_with_rxjs/creating_and_querying_observable_sequences/error_handling.html#ignoring-errors-with-onerrorresumenext https://xgrommx.github.io/rx-book/content/getting_started_wi...
PaulRobinson 10y ago
I feel sorry for the people who wrote this. Which one of us would want a code review by people who did not understand the requirements given by "the client" done on the front page of HN?
zobby 10y ago
I found all those rants amusing considering how many times nodeJs news reached top places on hackerNews.
lewisjoe 10y ago
And then I visited their homepage and found the UI to be uglier than the code.
skrowl 10y ago
While this code is pretty bad, I'm going to guess that the NSA reserves their best IT folks for... you know... the assignments other than their public facing website.
- pba 10y ago
  Agreed, this code was probably written by a contractor rather than one of their mathematicians I'd imagine
  - delinka 10y ago
    Have you seen mathematicians write code? For a website? We hired a few (we're in the crypto business) and we don't let them near production code. Ever.
    - sremani 10y ago
      General observation, mathematicians excel at functional v. imperative, partially because so many functional concepts are grounded in mathematics/logic.
vertex-four 10y ago
Honestly, it's... not terrible? It's hackish, sure, but the NSA website's frontend isn't a huge app where good software engineering needs to be practiced - the code does what it needs to do, was likely written quite quickly, and doesn't need the likes of webpack to build. In my core projects I'll use Ember or React+Redux, but where I just need a couple hundred lines of one-off code for e.g. a marketing site... why not some hackish jQuery?
Grue3 10y ago
A lot of popular websites have pretty crappily written JS code. The reason it's not noticed is that usually it's served compressed. The reason why it's crappy is that the bulk of functionality is implemented server-side, and the person who ends up writing that code is primarily a server-side programmer who half-asses a JS snippet to finish the feature, and then since it works, it is never refactored. It's not "real programming" anyway ;)
- FLUX-YOU 10y ago
  I would like to see a website with very good javascript code. I feel like it's basically impossible to write JS and have 10 or more people with 3-5+ years go "Yeah, I can't improve this at all and it's extremely clean".
  - Pr0ducer 10y ago
    About a year a ago, started from scratch writing a front-end client SPA for web-based software -- it's a custom framework built around Backbone.js and jQuery, written in CoffeeScript. The .coffee files are beautiful, and on save a shell script automatically compiles all the .coffee files down to a single JavaScript file that I would describe as good JavaScript code. Not open source, so no demo, but it's definitely possible to create good JS.
elcapitan 10y ago
They should add a comment ad in there at the bottom saying /* Smarter than that stuff? Come work for us. */
- borplk 10y ago
  /* Smart enough to not want to work for you sick bastards */

48 points by thewavelength 10y ago

maze-le 10y ago

> assets/js/main.js I smell rails

robin_reala 10y ago
Default Rails would be assets/javascripts/application.js
- _asummers 10y ago
  JavascriptS? Plural? Interesting.
  - clentaminator 10y ago
    Don't forget CSSes.
  - paulojreis 10y ago
    I also got intrigued when a fellow colleague used that nomenclature (folder called "javascripts", as well as "stylesheets", inside an "assets" folder).
tcfunk 10y ago
Is smelling rails a bad thing? (Genuine question)
- maze-le 10y ago
  Nope, I don't think so. I am a rails developer myself. It might have sounded a bit condescending...

DCRichards 10y ago

This isn't as secret as you might think... http://imgur.com/VYZbZKF http://imgur.com/VYZbZKF

mpeg 10y ago

// start comment block That's some properly documented code ! // end comment block

dagwaging 10y ago
my personal favorite: // form setup try { formSetup(); } catch(e) {}
0xfeba 10y ago
Heh. // title update try { titleUpdate(); } catch(e) {} // titleUpdate() //update title to include " - NSA.gov" on all subpages function titleUpdate() { ... }

cookiemonsta 10y ago

What would be the reason for this: // analytics disable if url contains underscore try { if(window.location.href.indexOf('_') > 0) {window['ga-disable-UA-20753548-1'] = true;} } catch(e) {} });

robotnoises 10y ago
My guess is that there is a test scenario in which they don't want google analytics to load?
mpeg 10y ago
It seems like they store files (presumably both internal and public) under https://www.nsa.gov/ia/_files/ https://www.nsa.gov/ia/_files/ so I guess it's the way they represent their non-indexable urls.
gizmo 10y ago
The if condition is wrong. This is why functions like indexOf should return a tuple of a boolean and an offset instead of encoding "not found" by returning -1.
- ChristianBundy 10y ago
  The if condition should use >= rather than >, but it otherwise makes sense.
  - dahart 10y ago
    I'm not arguing in favor of writing it that way, but I might argue it's not incorrect as written. location.href can never start with an underscore. A regex on location.pathname would sure be a lot clearer to read, though.

crisisactor 10y ago

Stop advertising the NSA please

BasDirks 10y ago
This surely isn't advertisement. The code is terrible. Try-catch jungle, ad-hoc URL manipulation, lack of caching, writing/constructing HTML in strange places, weak comparison operators in dangerous places (string vs lunch), etc.
- crisisactor 10y ago
  Yeah but when visiting in Lynx, it displays fine. Lynx is the yardstick I use to judge a website. There are countless .GOV sites like this made on old Windows 2000 machines with Frontpage installed, but my point is the information is still communicated. It probably doesn't matter about some spagetti JS lumped in. I assume any sensible person will have JS turned off when visiting a URL like that anyway...
  - BasDirks 10y ago
    I see. I think it's important to make websites work well for as many clients as possible, including lynx. Ideally, JavaScript should only be used for enhancement of necessary interactivity. But when a choice is made in favor of some technology, I like to see it used properly. In this case maintainability, performance and robustness could benefit from improving this code.
  - kiya0 10y ago
    :) man have you even met today's web developers? Turning off js is like asking them to work in assembly. Now that Google has gone and built everything the OS does, into the browser the mordern day web dev is programmed to think that they exist to rebuild desktop apps in the browser. Communicating information simply in plain text is the last thing on their minds.
    - BinaryIdiot 10y ago

zuck9 10y ago

Looks like it was written by a VB programmer, who's missing "on error resume next" in JS.

rbirkby 10y ago
Should have used RxJS https://xgrommx.github.io/rx-book/content/getting_started_with_rxjs/creating_and_querying_observable_sequences/error_handling.html#ignoring-errors-with-onerrorresumenext https://xgrommx.github.io/rx-book/content/getting_started_wi...

PaulRobinson 10y ago

I feel sorry for the people who wrote this. Which one of us would want a code review by people who did not understand the requirements given by "the client" done on the front page of HN?

zobby 10y ago

I found all those rants amusing considering how many times nodeJs news reached top places on hackerNews.

lewisjoe 10y ago

And then I visited their homepage and found the UI to be uglier than the code.

skrowl 10y ago

While this code is pretty bad, I'm going to guess that the NSA reserves their best IT folks for... you know... the assignments other than their public facing website.

pba 10y ago
Agreed, this code was probably written by a contractor rather than one of their mathematicians I'd imagine
- delinka 10y ago
  Have you seen mathematicians write code? For a website? We hired a few (we're in the crypto business) and we don't let them near production code. Ever.
  - sremani 10y ago
    General observation, mathematicians excel at functional v. imperative, partially because so many functional concepts are grounded in mathematics/logic.

vertex-four 10y ago

Honestly, it's... not terrible? It's hackish, sure, but the NSA website's frontend isn't a huge app where good software engineering needs to be practiced - the code does what it needs to do, was likely written quite quickly, and doesn't need the likes of webpack to build. In my core projects I'll use Ember or React+Redux, but where I just need a couple hundred lines of one-off code for e.g. a marketing site... why not some hackish jQuery?

Grue3 10y ago

A lot of popular websites have pretty crappily written JS code. The reason it's not noticed is that usually it's served compressed. The reason why it's crappy is that the bulk of functionality is implemented server-side, and the person who ends up writing that code is primarily a server-side programmer who half-asses a JS snippet to finish the feature, and then since it works, it is never refactored. It's not "real programming" anyway ;)

FLUX-YOU 10y ago
I would like to see a website with very good javascript code. I feel like it's basically impossible to write JS and have 10 or more people with 3-5+ years go "Yeah, I can't improve this at all and it's extremely clean".
- Pr0ducer 10y ago
  About a year a ago, started from scratch writing a front-end client SPA for web-based software -- it's a custom framework built around Backbone.js and jQuery, written in CoffeeScript. The .coffee files are beautiful, and on save a shell script automatically compiles all the .coffee files down to a single JavaScript file that I would describe as good JavaScript code. Not open source, so no demo, but it's definitely possible to create good JS.

elcapitan 10y ago

They should add a comment ad in there at the bottom saying /* Smarter than that stuff? Come work for us. */

borplk 10y ago
/* Smart enough to not want to work for you sick bastards */