My close friend and “big bro,” Bill Inmon, dropped an excellent article last week about Kimball vs. Inmon architecture. Bill recently moved to Substack and is republishing many of his old and new articles, and I believe this article might be an older one. I can’t speak to Bill’s motivation for publishing the article, but I’m glad he did. Any context or perspective from either Bill or anyone from that era is always welcome. I encourage people who contributed to our industry to record their experiences for future generations, especially people from Bill’s era.
The Kimball vs Inmon wars are the stuff of industry legend. To this day, they’re still spoken about in some circles. As expected, the article was catnip for the Inmon haters. A miniature Kimball vs Inmon turf war broke out in the article’s comment section. Some people have a weird habit of holding on to the past. Although Bill is a close friend of mine, I don’t have a dog in the fight of Kimball vs Inmon. It was before my time. To me, continuing to obsess about it is like arguing about which side had better trench warfare tactics in World War 1. While we can learn from those experiences, it’s also ancient history. We are in a far different place as an industry today.
The article and ensuing interest gave me the idea to put together a podcast discussion with Ralph and Bill. I wasn’t motivated to have a debate, so much as to host a discussion and a retrospective on their contributions to the history of our field. Ralph and Bill are both titans of the data industry, and I’m sure the discussion would’ve been widely watched.
Unfortunately, Ralph had no interest in participating in this discussion. I get the impression he checked out of the data industry long ago and has moved on. At the age of 81, I can’t blame him. As a side note, Ralph is alive and well, with a hobby of astral photography.
From what I understand (and this is coming from Bill), the war was not between Ralph and Bill, as they had no personal issues with each other. Instead, the war was between their disciples. Based on what I’ve been told, the core of the war was about the ownership of the term “data warehouse.” The secondary war was over which architecture and methodology was superior. Plenty of articles have been written about these comparisons (including Bill’s recent article), so I won’t rehash them here (Google is your friend).
A big reason I would’ve loved to talk to Ralph (or his colleagues) is to get his perspective. I’ve reached out to several people connected with Ralph, and have either heard nothing or received a decline. So, I can only go off of the things that I’ve heard from Bill and from other people who were in the thick of the drama at the time.
Some people accused Bill of pushing this podcast opportunity to “score points” against Ralph. I can assure you this is total bullshit. The podcast was solely pushed by me, with the motivation of giving the data industry a much-anticipated discussion. It would’ve been one for the ages. But to those claiming that Bill’s still obsessed with “winning” the war from several decades ago, please get a life. I know Bill well enough to know this is the last thing on his mind. He actually told me that after the data warehouse, he moved on to work on the gnarly problem of analyzing text data. The bigger point is simply that Bill and Ralph don’t care. If you’re still upset about the Kimball vs Inmon wars, see a therapist and get help.
A friend asked why there aren’t similar wars today. Back in the day, there weren’t as many big ideas in play, and Inmon and Kimball had the mindshare of practitioners and leaders. As a result, you might be expected to pick a side. Data identity politics was very real. Another big reason is that the news cycle was simply longer when the Kimball vs Inmon wars were raging. Reading back through old articles from the 1990s and 2000s, the data industry news cycle was whenever articles were published in print or online. Social media didn’t exist yet, so the reaction speed to these articles was much slower. Finally, there are so many rapid permutations of practices and technologies (new storage formats, AI, etc.) these days that it’s hard to imagine people foaming at the mouth over a particular issue. The identity politics of the past simply don’t manifest the way they used to. Even in the broader news cycle, it’s hard to get upset about a story for more than a day or two. There’s so much happening in the news - and our industry - that things simply move too fast. People forget whatever they’re mad about and move to the next thing, then the one after that.
In the end, the industry didn’t choose Kimball or Inmon. It chose both. For example, modern Lakehouse architectures can use Inmon-style governance for the raw/bronze layers and Kimball-style dimensional models for the gold/serving layers. The terms “data warehouse” and “star schema” are thrown around thousands of times a day around the world. The ‘winner’ wasn’t Bill or Ralph, but rather the synthesis of their ideas. Let’s give them their due respect and move on to building what’s next.