Although the development team deserves applause for introducing new features to support remote workers in the face of COVID and made improvements, the aggressive marketing of the numerous features in Nextcloud Hub 20 have aroused some of my suspicions. In fact, I am very concerned about the impact of cranking out major releases at a dizzying rate on home and medium-sized business users.
At a technical level, devices and network connections at home are frequently at a disadvantage compared to the office infrastructure, making performance a critical factor for home office workers. Under the hood, this release introduces a number of significant improvements:
Defining my concerns:
Although @JimmyKater believes that NC 20 is geared toward home users, I am hoping that developers would consider pushing back the end-of-life date for Nextcloud 18 and NC 19 to hopefully provide users more stability while the kinks in Nextcloud 20 Hub gets worked out.
With respect to those running Nextcloud with limited resources, e.g. those running the software on shared hosting, where certain errors are much more likely to result from bloated code, what do the developers have in mind for code review before the development process gets out of hand?
My perception is that too much time and effort is focused on rapid new release cycles and not bug fixes or stability.
Iām not saying that bug fixes donāt happen (eventually), just more that the priorities seem to be on rolling out new bells and whistles all the time.
I would overall agree with your criticism. I understand their longer term strategy (they need to massively expand what their software can do, and even if there are bugs, their most important [to them] users right now are paying enterprise customers who have a close contractual relationship with the devs who will get their bugs identified and ironed out quickly) but it really sucks for the community of people who are just trying to use it on their own in small office environments that canāt afford a multi thousand euro a year contract for services that other companies offer for like $120 a year.
First of all, new releases usually come with new features, often there are other major changes under the hood (that canāt be pushed in a minor release) that improve performance and stability (e.g. some versions back they did some major work on the database performance).
Furthermore, features are often requested by the community as well. Earlier there were some modules only for enterprise users, itās really bad because community users wanted this feature too. The latest development of having a full-blown suite of everything is requested by many users (āwhy doesnāt it run on my raspberry pi 2 properly?ā).
And the same thing for performance and bugs. Private users might have worse connections, in terms of performance, less load per user or per certain action would heavily benefit business users as well.
About the āmanyā bugs, I am not sure how many of those are related to specific environments or how many are a real problem for everybody. On the issue tracker you can also up- and down-vote issues, so it is possible to see a bit the spread of a specific issue. There is always the risk of specific issues being hidden by other bugs, feature requests and minor modifications. For that reason, Iād prefer to name specific problems. I havenāt followed many topics recently, I am not sure to ask obvious things, but if there is a huge amount of users that canāt upgrade to NC 20 because of these 5 major bugs in NC 20 (I hope they are not complaining at the same time in the thread why NC 20 is not pushed to the stable release channel). The setups of the community are very different ā¦
My main gripe with Nextcloud is actually related to the topic of this thread. Though Iāve not really had issues with bugs per se, but I do find that the team is adding features way too fast. The result is that even though a feature is released to the public and marketed, it is rarely useful in the first 1-2 years after release. Not due to bugs, but simply because of the starting feature set and usability. It happened with every major feature: Talk, Mail, the new Gallery, online Office, E2E, federation etc. Iāve honestly gotten used to not being hyped about new features because they arenāt really useful anyway (or at least very limited) when introduced at the start.
The usual response when something similar is brought to the teamās attention is that they need time for development because they have limited resources etc. Well, to that I can only say, Iām not surprised that development of every single feature is slow when theyāve got a dozen of immature features to work on at the same time. My advice for the team is to give more planning and design to new features and to bring them to a higher level of usability, maturity and feature set before introducing another major one. Similarly, after introducing a new feature, they should really focus on it a lot more in the months after its initial release to incorporate community feedback, before starting work on new shiny stuff.
Unfortunately, shared hosting is a bit of a surprise package when you want to use it with Nextcloud. Some are configured in a way that it is impossible to run Nextcloud properly. Especially for larger setups, itās getting complicated because of the runtime of scripts. Some hosters provide SSH logins that might solve this partly, adjustments on the database, redis etc. are usually impossible. I think ownCloud removed the support (or the install script) for shared hosting completely.
was announces as a concept and there was an early version which was still in alpha status. I suppose the idea was to get a lot of feedback on the design at an early stage but it wasnāt implemented as fast as announced.
Some parts (āhigh performance backendā) were closed source, now it should be better. Sure not very well documented, you have to search the detailsā¦
Was a development on the free time of some core developers and the community (slow work in progress). And there were other webmail implementations as well (rainloop, roundcube).
I donāt like this marketing presentation either. Iād prefer a new feature to be developed, perhaps introduced as alpha or beta version and after some testing put it to the list of official features. However, if you announce it then publicly, everyone will tell you that is a very well known feature used by many for some time. So it is less of a āsurpriseā and less of an interesting new thing. On the other side, companies looking for solutions compare theses lists of features, and the more boxes you tick the better, no question how good and reliable they are. For competitors only announcing well tested and very stable features, thatās a clear disadvantage.
Thank you, @rhua , for raising this important issue.
Most cloud services providers compete for the same business customers, and the larger such providers with always win, in the same way, in which Microsoft has won so many contracts, precisely with large organisations, despite the fact, that their product was the worst. Elephants feel always more comfortable in the company of other elephants.
As our NC will be anyway at least at a disadvantage (not to say a looser-in-waiting) in competing for such customers, it might indeed be an idea to focus on individual customers, like Bill Gates, ages ago went into another direction than IBM.
That it makes more business sense to go for individual customers has proven also Jeff Bezos.
For the future, being based in the Germany or Austria, with their privacy laws, will give NC anyway a clear advantage over being based in the USA or the UK. Precisely in this point I clearly disagree wtih @westernwashingtonit and his boundless understanding for businesses, so common among Anglo-Saxons.
In my opinion, the economy is there to serve the people; the people are not there merely to serve the economy. This seems to be a major difference in attitude between Anglo-Saxons and Europeans.
I wonder what others think of this, especially the NC decision makers.
Thank you all for considering this.