Last time I had a phone screening with the, I learned their entire codebase was in Perl, with no plans on migration to anything else. There were almost no unit tests anywhere. Most developers could push directly to production.
Everything the girl on the phone said was horrifying and I decided not to continue. Their recruiter kept trying to call me over Skype though, and they offer more than most other companies of that size in the EU, so I suspect they're pretty desperate.
This was back in 2015 so maybe it's changed, but it sounds like a terrible place as far as technology and tech debt goes.
> their entire codebase was in Perl, with no plans on migration to anything else. There were almost no unit tests anywhere. Most developers could push directly to production.
One of these things is not like the others. Not liking Perl is a valid personal choice, but doesn't say much about the codebase. The others are telling though. More so if it's true given that Perl has a very heavy culture of testing (the TAP protocol comes from Perl, and Perl is known for being very test heavy in its module ecossytem), which means they would be going against that culture to have no unit tests.
That said, I'm not sure if we should consider "the girl on the phone" doing the screening as knowledgeable about the process. That description doesn't convey that you thought very highly of her, so I'm not sure why we should take her word as authoritative...
The third hand knowledge I have of booking.com was that lack of unit tests was an explicit choice. I.e. they felt it was more efficient to just catch bugs in production and be quick to fix them than spend time in unit tests that got obsolete fast and add friction.
I didn't buy in the narrative, but it was an interesting perspective.
One should understand that at Booking, the only thing that counts is "conversion". In other words, the ability of converting a pageview into a reservation. Because that's the part of the process that creates value.
Since hotel reservations refer to a real world, where people change their plans, hotels get overbooked, burn down, are not reachable because of strikes, volcano eruptions, etc. etc. there is a huge customer service department. These people are able to cancel reservations, or force availability (after consultation with the hotel). In other words, until people have stayed at the hotel, and the hotel gets the bill for that, it really does not matter whether something was correct or not: the customer department can fix any problems.
So the only valid test is really: does it convert? And this is very closely monitored in close to real time.
The financial part of the system is thoroughly tested, as Booking.com is a listed company, so the scrutiny is extra severe.
> Most developers could push directly to production.
This is facilitates ownership and empowerment. You need to be sure that what you're doing, will not break things. If you are, it should be ok to push to production. If you were wrong, then that's ok the first time. Just don't do it again for the same reason: always make new mistakes!
This was the situation in early 2012, when I basically retired. Things have changed since then, but I still think the above is still basically true.
You’re not the first to mention Bookings lack of tests. Supposedly it’s because their code has a very short half-life, gets rewritten in no time and thus tests are seen as a hindrance.
My brother loves Perl, so when he was looking for a job, I recommended Booking. I don't think he ever even applied for a job there, though. Possibly because of all the dark patterns.
I still use booking.com quite frequently, but the dark patterns are starting to annoy me. I wish there was a better alternative tho, yes, I could book from hotels directly, but I still need to find those hotels, that's why I use something like booking.com. I've had hotels outright tell me to just book through a third-party booking website, since they didn't have their own booking form.
Best thing to do is find a hotel through booking and then book directly. Many hotels give you a discount for it too, because they don't need to pay Booking.com.
I recently had to book a hotel in Essen during the Spiel! game fair this week, during which nearly all hotels are sold out.
Using Booking.com, I could only find hotels that were too expensive or too far away. Using a different site, I found a really affordable one near the Oberhausen railway station, at half an hour from the Messe in Essen.
So Booking isn't even all that convenient. Yes, they have a lot, but they're also missing a lot. It pays to shop around.
I wonder about an app that just has phone numbers and a short description of hotels. If there are 50,000 hotels in the US, and you store 1K worth of data, that's only 50M. You could keep that on a phone no problem.
Do you frequently book hotels or travel? Just wondering because the app that you describe is so woefully short of even reasonable expectations, let alone what we have right now.
Once I checked Agoda to get an idea of the room prices in the area I was visiting, but decided to get the room directly from reception assuming that the price would be the one I saw on Agoda minus their commision. I was wrong, the price was about 15% more expensive at reception and it didn't include breakfast. I pointed to them that I could get the room cheaper from Agoda and they told me to get it from Agoda then. I sat in the lobby and booked the room online.
So you ring the hotel and the person who answers, if they’re worth their salt, tells you there’s only one room left / there’s a big convention on in town / everywhere else is likely fully booked out, and closes the sale.
Honestly it makes me want to use the website less, and does not urge me to buy.