[tombert]

I used to work for Jet.com, which was acquired by Walmart, and I stayed there for about two years after the acquisition.

It was an OK place to work, and they gave Jet some level of autonomy, but I left when I read that Walmart wants to start "integrating" their stuff ours, which in my case was a lot of Java crap that was clearly at least 10 years old.

I doubt that my experience is atypical with this. Big companies like to buy smaller ones, but don't like having two versions of one thing, and very often they'll view their original version of something as superior to the one they just purchased. In the case of Walmart, I think they bought Jet exclusively for the name.

I'm hoping that IBM is smart enough to sit back and let Redhat do their thing. I don't really want Redhat to become something bloated and horrible like WebSphere.

[kevstev]

Interestingly, I had a unique view on the other side of this. I worked at Walmart Labs at the time of the acquisition as well- remotely, from Jersey City. Long story short, upper management had changed, they were no longer supporting remote, at least not for management level, and I saw the writing on the wall that my future wasn't looking great- I fought against their massive tightly coupled Pangaea system (that Java code, shockingly was only a few years old, nowhere close to ten) and lost that battle, I was remote, my saving grace that I was working on grocery which was absolutely taking off at the time.

Anyway, I happen to go to a brewery and meet one of your product guys- he had a Jet shirt on, and I just went over and said hello. We got along great, famously actually, and after a few beers I said "hey you know, this remote thing... its not looking like a long term thing... you guys seem pretty interesting..." and we set up an informal meet and greet. I get along with everyone, and had a trip out HQ the next week. I poked around about Jet, and mentioned hey you know I live about ten minutes from their office, and yeah- all the talk was about how quickly they could get off their stack and onto Pangaea. I just fought that, and gone through the painful task of moving some stuff to their stack, and another 2 years of dealing with migrations and not focusing on customers just seemed terrible. I ended up leaving.

But yeah WMT in general is all about the "one right way" to do things. Which is great if its the right way, but I often used to say that there were a few bad teams that were just absolute anchors around the entire firm. I understand you don't want 10 teams building the same thing, but a little healthy competition is ok to me...

[emmanueloga_]

Where does Clojure fit in the technology stack of Walmart?

Cognitect has a few references that seem a bit old [1], but I see WalmartLabs release clj libraries from time to time [2], and I heard a podcast once where someone talked about working with Clojure remotely too [3].

1: https://www.cognitect.com/walmart-case-study.html

2: https://github.com/walmartlabs/lacinia

3: http://blog.cognitect.com/cognicast/087

[kevstev]

I knew these guys, IIRC, clojure originally came from an acquihired startup. This grew into a relatively small team working on an e-receipts project that was remote but largely based in Portland.

The most interesting thing about that team was that it appeared that not just a beard, but an epic beard was required to on it. I keep in touch with one of them, but he has left the firm. The others... I should probably send a note to- those were good guys.

I guess more generally though, they are "off strategy" and clojure was not making inroads at Labs, at least at the time I had left. There was a lot of gravity towards only using Java on the backend, part of my friction with upper management was pushing node.js as a backend technology. My feeling on the Clojure guy was that they were a good team so they were left alone since they weren't a core feature and they never generated any noise. That's just my take on it though, why after the big upper management change they really bristled against certain teams and technologies but were content with others was not clear to me.

Before I sent this off, I figured I should look at your links- the featured guy in the podcast was the one person whom I had mentioned that I keep in touch with but had left the firm- as have the other two that were on it. I recognize three of the contributors to lacinia.

[emmanueloga_]

Got it, that sounds kinda disappointing. Also, Jet was one of the bigger faces of F# when it comes to big companies using it, so I guess... double disappointment :-( (I'd love to see both F# and Clojure become more popular on the job market).

[vxNsr]

With the recent walmat redesign and the constant new website features, I was under the impression that Jet had totally taken over the website(walmart.com). It's interesting to see that the opposite was the case.

[kevstev]

Walmart has been making lots of e-commerce acquisitions, Jet was just the biggest. A lot of it is about growing outside of the traditional walmart customer- High end brands won't sell on walmart.com, and even if they did, high end shoppers don't go to walmart.com, but they have no problem going to Jet, or Moosejaw, etc. Jet was also framed as an acqui-hire of their management team, in particular Marc Lore. How true that is, I can't really say- another way to put it is, this seemed to be a move that came from above e-commerce.

The strategy is more about reaching different customers, and gobbling up potential competitors while they are small. The frontends may look a lot different, but its all one backend (or at least that is the end state for these acquired companies).

From my perspective, Walmart.com is finally seeing customer facing improvements as they have finally completed the backend re-write and migration which took literal years and held up most customer-facing work. Rewriting the FE, while not a small undertaking, is actually not that difficult to do with their architecture. Compare that to say Amazon, who I interviewed with, and they admitted that changing say the "Amazon orange"color to something else would be a herculean effort involving a year of planning across the entire company and lock step releases across the company. They seemed to think they had a better way of doing things, but after hearing that, I wasn't so sure I wanted to work there.

[geitir]

I feel with Amazon you can intrinsically tell by the formatting and lack of consistency in design between pages the hodgepodge of code running

[kevstev]

Indeed. And listen every business needs to make tradeoffs, especially when you are just getting started. The job was described to me as running the "front page of Amazon" which of course sounded very exciting. I went to final stages there and it became clear that I was really going to be running a content management system, which is a bit different than I thought the job would be, but fine. But my last on-site at HQ, they really seemed to have a tunnel vision view on the problem, and just how siloed each team was became clear. I forget the specific question they asked that made it clear that changing any color would be a herculean undertaking, but I remember just being annoyed at how they seemed to have this air of superiority that they did things so much better than others, and seemed dissatisfied with my answers- to the extent that at one point I felt the need to make it very clear that to change the Walmart logo consistently, or turn it from blue to yellow or whatever, would be a simple change I could push from my laptop and have live within an hour. They were implying I couldn't handle the complexity and scale of their site, and I was in disbelief at the mess they had made of it.

Honestly the whole recruiting process had turned me off. They had reached out to me, and they baited me with the run the "front page of amazon" hook. I knew things were not headed in a great direction where I was, and it was right before bonus/vesting, so while I was not yet actively looking, I said just have a talk, make some contacts there. The recruiter I spoke to was unlike any I had spoken to, he felt like he worked out of a call center. He made it clear the position was based out of Austin, and I wasn't looking to relocate. It also became clear to me that this guy was just trying to fill this role, he had zero interest in trying to see what else was open where I might fit. Even though I wasn't looking to relocate, I figured I could wow them and either get them to relent on relocation or find a more local role that would fit. 6 months and several rounds later I finally get an on-site. At this point though, I had started actively looking, and by the time I got on the plane, had 3 offers in hand. I wouldn't have gotten on the plane, but I did kind of want to see the process through, I had never been to Seattle before and had a friend there.

I was really disappointed in the whole thing really, particularly their recruiting process.

[DonHopkins]

Imagine the Java -vs- Objective C wars, if Apple had bought Sun.

[tombert]

Actually, Java is used fairly heavily at Apple for their backend systems. If you go on their job boards, you'll see quite a lot of Java positions.

[scirocco]

Everything Red Hat does is open source.

You can't just change the code - then you also change the whole product development model. If IBM wants a change - then it must be preferred by the upstream community in order for it to happen. If they bypass this, then they kill the existing model and will find most engineers leave