Comment by stephancoral | Hacker News Reader

Comment by stephancoral | Hacker News Reader

As someone who worked for the NYT until recently, I can tell you a major thing holding the Times back is the sheer amount of bloat and middle management in the business, especially tech, side of things. Useful, even revenue-generating projects are sidelined for political reasons (such as encroaching on another manager's 'territory') and months of work had been negated because a higher level 'architect' decided that only three languages can be used internally. Legacy projects are held on way past their sunset - that means literally months of dev time devoted to projects used by maybe a dozen people. Don't even get me started on the architectural inefficiencies that rack up tens of thousands of dollars a week to AWS (soon to be Google because the tech org loves to switch providers every year, also resulting in massive amounts of redundant work). The NYT online search is one area that is just bleeding cash - it requires a lot of machinery for very little. That's money that could be used to hire journalists and finance deep investigative dives.

The NYT does incredible work. It just needs to take a hard look at it's digital side and trim a lot of the fat. There's a lot of churn and little long term vision. Unfortunately layoffs / firings are rare in the tech org but buyouts and their ilk are widespread in the newsroom.

Is one of the three languages context still, or did they finally get rid of all that? ;)

(nytd 2008-2010)

Hahah thankfully the usage and number of DSLs has dropped significantly. I still had to write stuff using Drools (https://www.drools.org/) tho which was the worst thing I've ever had to do in my dev career

I believe the Holy Three are Python, Go, and Java. Personally the fact that Go is on there totally invalidates the list for me. It upset a lot of other people, too, because it's not very widely used in the org and is just a pet favorite of one of the higher ups. Was a serious source of friction.

Why? I understand the holy wars between devotees of Rust and Go, but if anything I'm impressed that the NYT would actually look beyond 1990s languages.

Because it wasn't a widely-used language in the part of the org the decree applied to. It was purely chosen to seem cool and trendy and also because it's one of the higher ups favorite languages.

Hello! I was one of those annoying "architects" saying no, but it wasn't because only three languages are allowed. When I left the Times about six months ago, systems were actively being developed in the following languages: PHP, Java, Python, Ruby, JavaScript, Go, Scala. (And ObjC and Swift, if you want to count those.) Plenty of unmaintained legacy systems in other languages, too. So proposals to add yet another language (Clojure, right?) were I think understandably received with some skepticism. It's not that it couldn't happen, just that there'd have to be a hell of a good reason, and there wasn't. It was unfortunate that months of work were invested before a proposal was written, but that wouldn't have been a very good reason to approve it.

For what it's worth, I agree with almost everything else you said.

Thank you for the clarification. I was told by my manager that new projects could only be written in the three above-mentioned languages so I was just going by that. And yes I was part of the Clojure lobby haha.

I really do hope they streamline a bit of the bloat in tech. It's a real shame. So many dev hours wasted in meetings. So many PMs who know more about the intricacies of Kanban and JIRA ticketing but nothing about the product. There's a lot of talent there, though. I think it can be done.

Oh man, count yourself lucky if you never had "fun" with context. :) For everyone who hasn't had that ... joy, imagine perl4, but compiled and running in its own proprietary vm with a ton of employer-environment-specific extensions. It was 100% in-house to the Times. When I interviewed, part of the process was them throwing a screenful of it at you and asking "OK, what does this do?" as a test of your general ability to reason out things you had no possible way of getting pre-exposure to. The only redeeming value that it had really was that it was a) pretty fast b) pretty small in terms of memory requirements. I can't think of a single thing about it that PHP wouldn't be a better fit for though on every metric, and I'm certainly no fan of PHP. After I left that job one of the biggest drivers for me wanting to take compilers eventually was thinking back to what a shitshow working with ctx was and how badly I would've wanted a ctx2php transpiler. I worked on the team that among other things handled the login/auth system, which was sufficiently antediluvian that it had a lot of ctx code.

Thanks for the input. What did you do at the Times?

I worked on cross platform services, essentially building out APIs and data delivery for both the newsroom and other tech orgs. I'm happy to answer any questions people might have about the Times - it's an interesting place to work (socially speaking, I found their tech to be pretty boring).

> I'm happy to answer any questions people might have about the Times

You asked for it. :) If any of these are inappropriate for you to answer, I understand. Thanks much in advance:

1) The Innovation Report came out in 2014. Per the article (and corroborated by other things I've read), it changed attitudes at the Times because "You couldn’t read that report and think that the status quo was an option." People there still thought the status quo was an option as late as 2014! Is it just standard incumbency psychology - where the incumbent power in an industry lives in an echo chamber, is tied down by vested interests, and can't see change coming - or is something else going on?

2) I've read a few internal reports, leaked and released. All repeat how great the NY Times is -'the best in the world', 'the greatest', etc. - over and over. It was in the Innovation Report, which was supposed to be only for top management to read. These reports aren't marketing to the public (and NYT marketing doesn't use that message nearly as much as the internal reports I saw) and these aren't stated missions ('our goal is to be the world leader in ...'); it looks like flattery in internal communications. It seems like an odd bit of narcissism. What about the culture requires them to tell themselves how great they are? Are they that insecure? Usually highly successful organizations don't need to say that to each other; they know. OTOH, maybe my sample size is too small and it's not really an issue.

3) It still seems to me that they are still a newspaper (emphasis on that old medium, paper) making accommodations to the electronic medium, rather than a news publication on the Internet. Stories still are told almost entirely in text, which was a limitation of the old medium but not the new, and images and video usually are decoration rather than a primary means of story-telling. Even two-bit bloggers will, rather than describing something, just show it to the reader: 'The building is rundown: <img>'. I read a story that described how some innovative high school dance program started class each day with (a dance and song?) - why not just show me a video? Occasionally they have helpful graphics, but they talk like that's a special event rather than a norm of the medium and usually it's a separate story - 'the graphics story' - they can't seem to tell a story integrating different mediums as needed. What is going on? Are they just stuck? Will they let go of the old medium before it's too late?

EDIT: A few edits

> Stories still are told almost entirely in text, which was a limitation of the old medium but not the new, and images and video usually are decoration rather than a primary means of story-telling.

I much prefer the largely text medium, even in the digital age. One can consume & process content immensely faster with text than inages and video.

> One can consume & process content immensely faster with text than inages and video.

I prefer whatever is more efficient too, and text is great. For example, generally speaking, I prefer text to a video that is longer than maybe 30 seconds, because reading is random access and I can skip to what I want.

But as a universal statement, I'm pretty sure the above comment is factually inaccurate. Certainly for large classes of information, visuals are more efficient. For example, a graph of year-by-year unemployment rates in the 20th century is far more efficient than a list of numbers and years. Similarly, a photo can convey information that would take pages and pages of text.

A photo leaves too much interpretation to the user, which takes away from what you want from the news - information.

Graphs are also extremely specific, and very easy for people to misinterpret/get misled by. The specificity of graphs also means that most forms of information cannot be conveyed that way, and so it is a limited tool.

Reply on news.ycombinator.com