I have been an enthusiastic supporter of Open edX since it was launched and have pushed for its use in my country (Norway). At the time we used our national service “Sikt” for public Open edX implemntation while maintaining local servers for testing, drafting and backup/revisjon of course titles. The Tutor development was very promising. This has been sold as a “simple” and “one-click” approach.
The “tutor local quickstart” command is, of courrse, a good example of this. But having installed each and every revision since 2016 I must say that the previous solution from Bitnami IN SUM has been easier. The reason is that it has been consistent. With Tutor I never know.
Now I am running two different instances of Maple 2.2 on two Ubuntu servers. The installation procedures seemed to be identical to me.
It is supposed to be simple. But one is running (not on Google Chrome, though. I have to use the Beta version). The other is not responding with a blank page when I try to open a course title. I have not got the faintest clue as to why.
I have been in the field for more years than I care to remember (good ol’ Fortran days…) but I am not well trained in Python and Django. But do you have to be a recent computer engineer graduate to participate in a meaningful way? That is the opposite of “simple”.
If this is not rectified with the Nutmeg release, I am afraid - and unhappy - that this project will not fly. What are the future prospects for Open edX?
I have been thinking about your comment over the weekend and I honestly don’t know what to answer. Are you saying that running Open edX is hard? I agree – believe me, everybody does. Are you saying that Open edX should have stuck with the native installation? There was a long conversation about that and there was unanimous agreement that Tutor was the better alternative. Do you have something concrete to suggest or are you just ranting?
In what sense? I don’t even know where to start. If you would like to discuss the future of Open edX, I suggest that you join the conversation in one of the many working groups.
I think there is a mismatch between the real complexities of installing and maintaining the Tutor solution for Docker and the professed simplicity of “1-click installation” with the accompanying professed goal of making it “easy to deploy, customise, upgrade and scale Open edX.”
You state that “Tutor is reliable, fast, extensible, and it is already used to deploy hundreds of Open edX platforms around the world.” I find this to be a bit of an “oversell”, even with those hundreds of installations.
I think the concept is very good and I understand and respect all the work that is put into this.
Waiting for the Nutmeg release I hope that a standard installation will work in the professed way with standard browsers like Google on standard Windows machines. If not, it would be better with some warning and an indication of which release is stable.
I do not want to revert to installation procedures before Tutor. But the Bitnami-based solutions were well documented and I knew what to expect when I tried to install a new system based on them.
It seems to me that maintenance “hangs” on one individual. Is this correct?
(Context: I am trying to promote the system to a community of tens of thousands of employees and hundreds of thousands of students in Higher Ed in Norway where Canvas is the dominant solution.)
I am sorry that Tutor does not work for you – but really, Tutor does work very well for many people. I am open to constructive criticism to make Tutor and Open edX better. If you are part of a large organization with consequential funding and engineering resources, and if you really want to use Open edX and Tutor, then it is expected from you that you contribute back to the project, either financially or with your technical expertise. This is just how open source works.
The standard installation does “work in the professed way”. Maple is the current stable release. Open edX does run correctly on Chrome on standard Windows machines, as far as I know. If one of your platforms works and the other does not, then you should know that there must be some sort of difference between them. As an engineer, it is your job to report the issue in a professional manner.
This is only partly correct. As a BDFL I act as a central point to manage and coordinate the project, but maintainers do contribute code and respond to forum queries. I would not be able to keep supporting the project without the rest of the community.
I believe that you and we need to address this differently. I have no problem in understanding that a complex piece of software will require attention and resources by those who engage with it. The problem, as far as I can see, is this: In the communication about “simple 1-click” installation is not entirely correct. It misleads people to think that this is more simple than it is.
I just did a clean re-installation of Maple 13.2.2 (for the umpteenth time). I seem to follow each and every instruction given. The Studio part works well. The display of available courses works well. But I get blank page or a “Error loading…” dependent on which browser I use when I want to open a particular course title.
It seems to me that version 11.1.11 is more stable.
Would it be possible to identify 5-10 SysOps around the world that are willing to do clean installation tests of any new release for one or more environments before this new release is made public? Maybe most of the perceived problems could be sorted out early on in a systematic fashion? How do you go about this kind of “user testing” now?
I hope that I do not come across as quarrelsome. I really appreciate all the efforts and am a great believer in Open Source and Open edX.