Why (A)GPL software is better for your business - give input please!


  • Vendor

    I recently published a blog about open source licenses.

    Now isn't hard to argue when proprietary licenses are so universally shitty - I link to this and it is a good read.

    But of course there is more to it. Certainty about longevity of the support, quality - they are arguments, too. You still have to do your own investigation, actually I often see business people don't realize this nor know how easy it is.

    So a special tip for all of you: want to know how an open source project is doing? Check out Github Pulse and other statistics!

    As an exercise, and note I obviously have a stake here and am totally biassed about file sync and share, but here are four projects. Compare and tell me which one is doing great and which one is closer to death:

    https://github.com/nextcloud/server/pulse
    https://github.com/pydio/pydio-core/pulse
    https://github.com/haiwen/seafile/pulse
    https://github.com/owncloud/core/pulse

    Be sure to digg a little deeper. Sometimes, the monthly view looks different. And there are more statistics behind the pulse, which come with developments over time.

    You can see, for example, that pydio has a very cyclic development process here:
    https://github.com/pydio/pydio-core/graphs/commit-activity
    0_1478790518038_Spectacle.b13983.png
    While Nextcloud is more variable:
    0_1478790628734_Spectacle.n13983.png
    And ownCloud is on the way down:
    0_1478790653926_Spectacle.R13983.png

    Note also that each project has more repositories, though these core repo's are often the most important ones, they don't tell the whole story and sadly as far as I know, github doesn't provide a real overview of an entire project.

    So, when your boss asks you to pick a project or product to use, this should be a part of it: how healthy is development!



  • @jospoortvliet This is actually one of the reasons I tell people to hold off on forked projects until they have been running for about a year.

    Development always spikes for the fork and sags for the original. But after a bit of time you will see if the original rebounds or languishes as ownCloud is doing.



  • Also important to note that the scale on NextCloud and ownCloud on the graph are the same, but Pydio's scale is 33% that of the other too. So it appears much busier than it actually is.


  • Vendor

    @JaredBusch I agree, it makes sense to wait and see - we've been telling people that, too. Though I wouldn't wait a full year - a year is a loooooong time in IT. It's been 6 months and no sign of rebound of oC nor any sign of us slowing down... 😉

    But sure, wait until Nextcloud 11 is out, I'd say. About one more month and you'll have a number of clear feature improvements to pick reasons to switch from!



  • @jospoortvliet yeah year is probably too long. But I have yet to actually convert a existing system from OC to NC.


  • Vendor

    @JaredBusch out of curiosity, what would make you (or customers) switch? Is security a real issue, stuff like this for example?



  • @jospoortvliet said in Why (A)GPL software is better for your business - give input please!:

    @JaredBusch out of curiosity, what would make you (or customers) switch? Is security a real issue, stuff like this for example?

    At this point it is the need for testing.

    The problem is that one client has about 400gb of data in their ownCloud over something like 1 million files. I will have to kill replication for this VM in order to use the space currently consumed by the replica to test the upgrade.

    Also, I have not seen many examples of people upgrading large instances. I am sure people have done it, but I have not read about it.

    My understanding it to go from oc 9 to nc 9 then upgrade to nc 10. This will be very time intensive. Thus, expensive.


  • Vendor

    @JaredBusch said in Why (A)GPL software is better for your business - give input please!:

    @jospoortvliet said in Why (A)GPL software is better for your business - give input please!:

    @JaredBusch out of curiosity, what would make you (or customers) switch? Is security a real issue, stuff like this for example?

    At this point it is the need for testing.

    The problem is that one client has about 400gb of data in their ownCloud over something like 1 million files. I will have to kill replication for this VM in order to use the space currently consumed by the replica to test the upgrade.

    Also, I have not seen many examples of people upgrading large instances. I am sure people have done it, but I have not read about it.

    My understanding it to go from oc 9 to nc 9 then upgrade to nc 10. This will be very time intensive. Thus, expensive.

    As oC 9.0 = Nc 9.0 (with some fixes) you can upgrade directly to Nextcloud 10 (oC 9.1 + some extra features). But, like any upgrade, yeah, you should test and that will indeed cost time.

    With regards to big upgrades, a lot of providers have moved over loads of customers and we've migrated a bunch of customers as well but I'll see if we can do something public about that. It feels like insane to do a white paper about it as migration is just a normal upgrade but perhaps it helps convince some people...



  • @jospoortvliet a white paper may be overkill, but a blog post talking about some details of users counts and size counts would probably help to convince people.

    Again I, am already ready to migrate, I jut need testing time.


  • Vendor

    well I can talk to a service provider who has migrated, a few have - but we simply haven't done any huge migrations yet. Most of that is because big installations usually move slow - nobody wants to be first. Not that upgrading is a big deal, it is the same as an upgrade of course. In a way, once ownCloud does a new release customers should be more worried about that than about upgrading to Nextcloud... But yeah, how to explain that 😉

    We will see what Nextcloud 11 brings, people might find compelling reasons to upgrade at that point.


Log in to reply