Emils Solmanis

the bad side of apache

If you've met me at a pub and ever asked me what I think about the Apache Software Foundation (ASF), you've probably heard a somewhat constructive rant and an assortment of conflicting thoughts. I'm not alone in this, many other devs I know who've followed projects more closely, e.g., by reading the mailing lists, have similar thoughts. Today I'll try to explain why I think the ASF is too big and stagnant for its own good based on a series of events unfolding in the Cassandra community, and why places like GitHub have flourished.

at first, there was data(stax)

To give some background, Cassandra is not new. It's more than 8 years old now, if you count from the date of it first getting open-sourced. While it was initially made by Facebook, for the last couple of years, the driving force behind its development efforts has been DataStax, a consultancy built around the database.

Apache projects are lead by a Project Management Committee (PMC), which is basically a group of people who make decisions about the project's general direction (and other more detailed matters). The PMC has a chairperson.

Apache itself has a board of directors, which consists of nine people.

Under normal circumstances, the PMC's chair produces quarterly reports for the board and runs most communications between the two groups.

when the board gets bored

It all began on June 4th, 2016, when a wild Chris Mattman showed up, asking about whether DataStax "controls" the driver code. Whilst the question itself is valid, it's the bit where Mr. Mattman wouldn't take "no" for an answer that proved to be painful. He also requested a special report with answers to specific questions, but since one was due in August anyway, Jonathan Ellis, the PMC chair, said on the thread that they'll be included there. Having observed Apache's religious nonsense before, I knew the direction this was headed in, but I'd hoped there'd be the usual back and forth of passive aggressive messages, before the rest of the board finally reined him in and that would be the end of it.

silence of the boards

For quite some time there was silence on the public lists about this, and I thought this had died off. However, come August's board meeting, the minutes say that

The Board expressed continuing concern that the PMC was not acting independently and that one company had undue influence over the project.

PMC representatives acknowleged the Board's concerns and asked for clarity from the Board in the form of explicit actions that the PMC should be taking to improve the situation.

It was suggested that a representative from the board could meet with senior company management to relate the board's concerns directly.

It was agreed that the project should report again next month.

Two days later, Jonathan Ellis "steps down" as chair.

Whatever happened between the first thread in June, the board meeting in August, and Jonathan getting kicked off is all on private lists, despite Apache's ideals of open communication. Nonetheless, some of the communication has been divulged to the dev mailing list because many members thought the conduct of certain members of the board deserved public recognition, most notably threats to "jettison all DataStax employees from the project" and childish links to clips from Game of Thrones.

the downwards spiral

After the events in August, there was silence again until the beginning of November, when DataStax announced that they're refocussing on their enterprise edition product in the future. While Jonathan has assured the community that almost nothing will change, if you actually read between the lines, this is DataStax cutting it's losses as a business and giving the finger to the ASF, which most likely means new feature work going into DSE instead of core, and core only getting bugfixes. Given DataStax's role in the project, this is likely to stifle work for months to come.

Furthermore, when the community expressed concerns about all this, it seems nothing has changed in the board's attitude and manner of communication since August. Chris Mattman used ad-hominem attacks when called out on his behaviour, and when called out on it again, claimed that it's just the people misunderstanding him.

As of the moment of this writing, there has yet to be any serious acknowledgement of misconduct from the board, much less an apology or any concrete actions.

And this, folks, is why many of my dev friends have a strong dislike, if not hate, for Apache. This is by far not the first similar incident, but I'll let you ask Google if you want more.

Let this, therefore, be another call for Chris Mattman to be sacked from the Apache board. Conduct like this should not be tolerated for one of only nine highest ranking people in the ASF. If this is the face of one of (if not the) largest OSS organisations in the world, why would anyone ever want to work with it or endorse it?

If you enjoyed the read, drop us a comment below or share the article, follow us on Twitter or subscribe to our #MetaBeers newsletter. Before you go, grab a PDF of the article, and let us know if it's time we worked together.

blog comments powered by Disqus