It is funny this kind of came full circle. Damien Katz, the original CouchDB author was working on Lotus Notes as some point in his life. After many years, his "creation" comes back to IBM. There is some irony in there I think.
I remember him disavowing CouchDB and urging everyone to "ride with him" towards Couchbase. There were promises about how CouchDB will get contributions from Couchbase, and how it will get RPM and packaging scripts, docs and continued help. But not much of that happened. The community split. Most CouchDB developers who ended up on the other side of the fense at Couchbase didn't end up committing or helping that much over the years. So I can't see, looking back, anything good at all about that split.
But Cloudant sort of took over helping and watching over CouchDB. Quite a few of the contributors work for CouchDB, and they are very smart and helpful individuals. I hope this means good news for CouchDB as a project. I would want to see it grow.
It is funny this kind of came full circle. Damien Katz, the original CouchDB author was working on Lotus Notes as some point in his life
It's definitely interesting. I can remember people in the Notes community wondering why IBM didn't either take CouchDB and use it as a new Notes backend, or alternatively push the existing Notes bits as a mature document-based database.
The end quote from Katz's "Formula Engine Rewrite" post[1] is interesting:
"But ultimately IBM killed Iris. In late 2001 IBM folded Iris Associates and converted all the employees to IBM. All the energy the Iris building had was seeping out. Really it was more pushed out by the BS that IBM management kept trying to sell us..."
Damien's writeup is well worth a read, esp. by programmers who want to know what it takes to get to a high level of proficiency. The Iris work appears to have been a critical experience in preparing Damien to write Couch. Anyone who wants to become a great programmer should read this and emulate his example.
I was at Iris around that time too, and went through the IBM slow-death. I really felt like we had to spend at least one whole day per week (maybe more) dicking around with HR stuff. Career objectives, self-eval, performance reviews, peer reviews.
Facilities and IT systems were a horror show too. I spent 2 whole days registering with various internal websites, filling out forms, getting permissions - just trying to get rid of an unwanted monitor. At Iris you'd just put unwanted stuff outside your office and it would be gone the next day.
Funny, losing the amazing Iris tech-bubble perks kind of sucked, but it was really the daily "being an employee" overhead and red-tape that made IBM unbearable.
> It is funny this kind of came full circle. Damien Katz, the original CouchDB author was working on Lotus Notes as some point in his life. After many years, his "creation" comes back to IBM. There is some irony in there I think.
At Couchbase we created two open source mobile embedded databases (iOS and Android) that can sync with Apache CouchDB / Cloudant / PouchDB. Which is no small contribution. Even Cloudant bases thier mobile stuff on that code.
I've been working with CBLite for iOS and it's pretty legit. One thing that I have not been able to find on Couchbase's website is a written commitment that CBLite will always be compatible for CouchDB. I know that the current sync gateway behaves similar to CouchDB and so there is an incentive for interoperability, but it would be cool to know you guys will always have CouchDB users' backs on the mobile front so that we can trust you and continue to rely on your code. Who knows, maybe that would entice me to move to Couchbase in the future!
Thanks for the feedback. It's not my place to make commitments for the company, but I can tell you an engineer's answer:
I've yet to see a logical model for multi-master replicated documents that beats what Damien chose when he created CouchDB. So I'd be very surprised if that changed, and as long as that stays the same, interop at the transport layer isn't hard to maintain.
However, there are lots of ways the transport / wire-protocol that we use to transmit the logical model could be improved. So in my perfect world there would be many alternative transports, and the various couches could implement pluggable handlers for them. In the end this would preserve interoperability but also make room to experiment with the details. If you are interested in those conversations they are happening here: http://mail-archives.apache.org/mod_mbox/couchdb-replication...
I don't think it's irony. Early jobs in your career drive your perspective on technology and culture. That imprint stays with you, and you build new things in that model. Even if it wasn't explicit, the child mimics the parent. In some cases it is explicit. (I'll make something that my old employer will want to buy, and configure it to be as obvious as possible a decision)
I remember him disavowing CouchDB and urging everyone to "ride with him" towards Couchbase. There were promises about how CouchDB will get contributions from Couchbase, and how it will get RPM and packaging scripts, docs and continued help. But not much of that happened. The community split. Most CouchDB developers who ended up on the other side of the fense at Couchbase didn't end up committing or helping that much over the years. So I can't see, looking back, anything good at all about that split.
But Cloudant sort of took over helping and watching over CouchDB. Quite a few of the contributors work for CouchDB, and they are very smart and helpful individuals. I hope this means good news for CouchDB as a project. I would want to see it grow.