Feedback requested: when to migrate to composr.app

Considering options for when we should migrate to composr.app. Feedback appreciated!

Hello everyone!

As part of Composr's shift to v11, we will be migrating compo.sr to composr.app at some point. The question is when, which is what I would like your feedback on.

Current Plan

The current plan is not to migrate compo.sr to composr.app until v11 reaches its first Release Candidate status (e.g. a step just before stable).

Benefits:

• The site will be in a more stable and tidy state once the migration takes place.
• This allows the services offered by compo.sr to run a bit longer.

Cons:

• We're currently stuck in a "limbo" between the old support model (former ocProducts, Ltd.) and the new one for v11 (Bazaar model; heavy community support / ecosystem). This plan will mean we are stuck in limbo a longer period of time, which may be detrimental for fostering an environment for others to contribute to the project. Essentially, there is no support structure right now as we await v11.
• I do not know when v11 will reach RC status. It could be a lot longer than originally planned given there are still some bugs that need fixed which I alone am struggling to fix.

Proposed Alternate Plan

We migrate compo.sr to composr.app when v11's upgrader is stable enough to facilitate the migration even if the rest of the software is still in beta.

Benefits:

• We will be on composr.app sooner, which means the new Bazaar model, constitution, ecosystem, etc. can take effect. This will IMO better facilitate community support and development as we will have a structure again, therefore will also benefit the remainder of v11 development.
• The migration date will be less influenced by delays in v11 development.
• v11 will get more live testing, so we can find and fix more bugs (Mythus pointed this out in a topic further down)
   

Cons:

• composr.app may still be a bit unstable / rough when the migration takes place. I don't anticipate anything critical will happen (and I run a three-tier backup system in case something does). But there could be some quirks.
• compo.sr will become read-only before v11 reaches stable once the migration happens. While this won't affect the ability to download / upgrade v10, it may be slightly confusing at first (although redirects will be put in place on compo.sr).

Why can't we just put the Bazaar model on compo.sr?

The new support model utilises new features that are present in v11 but not v10 (e.g. voting power on topic polls, and the tracker utilises points for sponsorships in v11). These cannot easily be migrated to v10 (and would fall outside of v10's "Long-Term Maintenance" definition as well, so it won't happen regardless as I wouldn't have the funding to make it happen).

---

Please let me know what you think we should do. Feel free to also suggest other ideas I did not consider.
 

As v10 is the current stable version, it probably makes sense to stick with the Current Plan otherwise the v11 forum will potentially have topics concerning v10. Also I think it might become confusing in other ways if the v11 website isn't finalised and we are using the v10 tracker for filing issues.

Actually, that is a good point. I need to think about how we're going to handle the forums.

My idea is that all v10 forums and topics will be dropped into a top-level "compo.sr archive" forum on composr.app to easily separate it from the v11 ones. I definitely want to migrate them over because some v10 topics can still be useful to v11. And there are some general non-v10 / non-Composr topics that could still apply.

When we migrate to composr.app, the tracker will also migrate there. Ideally I'd also migrate all of the issues and comments there as well.

There would be redirects, so when the migration happens, if you try to go to the forums on compo.sr, it will redirect to the compo.sr archive section of the composr.app forums. And if you try to go to the compo.sr tracker, it will redirect to the composr.app tracker. Etc.

Probably the only sections that would not have redirects on compo.sr are the tutorials (v10 tutorials should still be viewable for those still on v10), downloads (so you can still download v10), and anything v10 specific which would not apply to v11 nor be on composr.app.

My thoughts are to go ahead and transition now. Here's why:

You'll catch more of the bugs in a live environment.

Yeah, I know it is not an ideal situation, but it is a true statement. Apart from me and a handful of others v11 isn't getting as much testing as it probably should. Composr already has a small footprint in the cms market share as is. V11 can benefit greatly seeing how all the functionalities work in a live environment.

I have recently made this move myself. Apart from the nonworking redirect function, it functions well enough for me to run it live for my purposes. But my purposes are not, at the present, large scale usage of the site. It is primarily the wiki and catalogues lol. I won't get the types of bugs you might find in a live community environment.

This isn't unprecedented either. Various other community software have ran live in beta. One more or less well known example is SMF. They ran their forums for their community in beta. I believe it was of great benefit to them as well.

I would suggest keeping a backup site running however. You could keep compo.sr running as a backup site, just change the domain or something so people don't just go to it. Then if the composr.app site has a catastrophic failure, which is not very likely at this stage, you could lock it with a redirect to the backup site while you fix the main site.

This would also showcase to the world the coming benefits of the v11 upgrade, which could generate excitement and maybe more involvement.

Help and support can still be given for v10 on a v11 site. The v10 tutorials can still be hosted. There can be a dedicated v10 support forum. There really is no loss for v10 support by going to v11 for the main.

"Eat your own dog food" is the philosophy here. v11 is where the majority of the resources are being spent currently. v10 for all intents and purposes is simply in containment mode. It is time to let go of the past and step boldly in to the future. It is time for composr.app.

Mythus said

My thoughts are to go ahead and transition now. Here's why:
You'll catch more of the bugs in a live environment.

That is a really good point. Composr is a Composr (community) site, not a business site. It doesn't have to be stable in the sense of completely bug-free, just functional and free from critical issues. And you would be right; running v11 in a bigger production scenario such as Composr's homesite would expose it to more potential bugs which the community and I have not yet caught.

And the upcoming beta5 should, in theory, be the most stable v11 release to date because I've fixed a lot of issues with errors being thrown when you have certain addons uninstalled. I also fixed issues pertaining to sorting. And I'm slowly but surely knocking out some of the bugs which have been reported on the tracker recently.

v11 is still not where I want it to be stability-wise to call it RC (release candidate) though. And it probably won't be until I can receive the help of Chris on a few issues I've struggled with. But using it in a live environment as the homesite could get us closer to that.

Mythus said

I have recently made this move myself. Apart from the nonworking redirect function, it functions well enough for me to run it live for my purposes.

I'm in the same boat. As much as I like v10, I stopped using it. All of my sites are now v11; I personally like it a lot more than v10 even though v10 was also great. My business site is v11. My pinball project site is v11. I have three clients in which I run a combined 4 v11 sites. So that's a total of 6 v11 sites on my end right now. Count composr.app since that's technically also a "client" website (since Chris is a client) and that's 7.

My two biggest reasons for running v11 instead of v10 despite it being beta: it looks better (and is mobile responsive), and it is more developer-friendly.

Mythus said

This isn't unprecedented either. Various other community software have ran live in beta. One more or less well known example is SMF. They ran their forums for their community in beta. I believe it was of great benefit to them as well.

That's true; I hadn't thought about that before. It's not unheard of in the Open Source realm for developers to run their own software live in the beta stage considering the philosophy of Open Source largely emphasises collaboration (and thus, collaborative bug-hunting and testing).

Mythus said

I would suggest keeping a backup site running however.

That was my plan all along; compo.sr would still be online for some time after the migration to composr.app. For how long, I don't know. That's up to Chris since he runs compo.sr (and I run composr.app). Actually, it needs to stay up so that v10 sites can receive the v11 upgrade via the Admin Zone.

Mythus said

Then if the composr.app site has a catastrophic failure, which is not very likely at this stage, you could lock it with a redirect to the backup site while you fix the main site.

It is extremely, extremely unlikely. Even if a catastrophic failure does happen, I have backups of composr.app, so the site can be reverted to an earlier point. My server backs itself up every day. And I download those backups to my local machine as well so I have an on-site and off-site copy. The server also runs in RAID so it can keep running if a hard disk fails (so long as it's just one failure and not both at the same time).

And even if all that fails, at least the core site code and content is on GitLab (database data isn't, like user accounts and whatnot, so that would still not be an ideal situation).

I take redundancy very seriously ;)

Mythus said

Help and support can still be given for v10 on a v11 site. The v10 tutorials can still be hosted. There can be a dedicated v10 support forum. There really is no loss for v10 support by going to v11 for the main.

Actually, the v10 tutorials cannot be hosted (on composr.app). With the way Composr handles tutorials, my server isn't able to handle tutorials for multiple versions without getting confused. I already tried that lol. But my plan is compo.sr will stay up for a while where the v10 tutorials will be hosted. And composr.app hosts the v11 tutorials.

I've learned that my server has a lot of quirks compared to Chris' which require me to take alternative approaches. For example, my server has tighter security, so I cannot host Git repositories directly (the Control Panel manages the Git repository rather than it being available in, say, the website directory for use with the git command). This breaks the maintenance status page which used to use the git command for checking the status of branches / versions of Composr. Now I have to use the GitLab API for that. I also cannot compile Composr releases directly on my server for the same reason.

My server also has the dreaded ModSecurity. Well, I can always turn it off, but I don't want to do that. It's actually a very handy tool for security when configured properly. I just have to be very mindful of what composr.app tries to do so it doesn't trigger ModSecurity.

Also composr.app requires more PHP children than usual for some reason… all other v11 sites work fine with the default 10, but composr.app needs more child processes. Go figure…

Mythus said

"Eat your own dog food" is the philosophy here. v11 is where the majority of the resources are being spent currently. v10 for all intents and purposes is simply in containment mode. It is time to let go of the past and step boldly in to the future. It is time for composr.app.

You should be the marketing director of Composr with that pitch at the end LMFAO.
 

---

Regardless of what we decide, the migration to composr.app still depends on some input from the community first. I need to finalise the new documents I wrote, especially the constitution. I do not trust myself to write perfect documents. I think you already gave your input, at least on the constitution.

I also need to tidy up the site more because it's a bit ugly. And I need to make extra extra sure the Composr import tool and the v10 -> v11 upgrader works as it should so that compo.sr can be migrated. I can't just simply upgrade compo.sr and use that because I already have a lot of material on composr.app. And I can't simply use the importer because you can only import Composr sites of the same version. It's a great opportunity to fix bugs in them, though.
 

LOL I was an evangelist for the Church of ocPortal at one time!

mythus said

LOL I was an evangelist for the Church of ocPortal at one time!

From “Post #10,177”, 15th November 2024, 11:38 pm

HAHA so was I. How far back do you go? I only go as far back as v9. (This will be my only off-topic question, lol)

I came in at the very end of the ocPortal 3 series right at the transition to ocPortal 4. I have experienced some rough transitions lol.

Wow, you are an old timer, lol.

What I find interesting is I just noticed… v11 has been in development longer than the time frame between the release of ocPortal v1 and ocPortal v9 (8 years).

I agree with Mythus, its time to end confusion and boldly go where no man has gone. While version 10 is stable it also contains a bunch of very old content, missing links and add-ons in the tutorials (IE Theme Reset that someone wrote but now is not found or is of any use). A lot of references are from years ago and programming has changed. What worked then doesn't apply now and creates confusion. What are we trying to push V11 or V10?

I see this situation like translation V10 (Spanish) V11 (English) and we have a translator that points to all things in both languages in mixed content causing people headaches. Get used to speaking Spanish then told you now speak English but use Spanish references from 10 years ago to mix the two. My gosh, I am getting a headache because I speak Rat and want a site that's easy to maintain and use.

Master Rat said

I see this situation like translation V10 (Spanish) V11 (English) and we have a translator that points to all things in both languages in mixed content causing people headaches. Get used to speaking Spanish then told you now speak English but use Spanish references from 10 years ago to mix the two.

I'm not sure I follow what you are talking about. Can you elaborate more concisely?

Master Rat said

A lot of references are from years ago and programming has changed. What worked then doesn't apply now and creates confusion. What are we trying to push V11 or V10?

To be honest, even the v11 tutorials need public review by people when they have time. I do not have the time myself to go through all of them for accuracy when I need to focus on v11 development. So the v11 tutorials may also have outdated information until people can help me sniff that out and correct it. Not to mention, a majority of the v10 and v11 tutorials were written by Chris. And he has a vastly different writing style from me. He's more elaborate whereas I want to stick to the main points in a very structured way (concise).

I also want to make more video tutorials, but that's a big time commitment.

Point is, I can't do everything. And Chris' goal (and mine as well) for composr.app is for both of us to step back a bit and let the community take more proactive roles.

Quote "I'm not sure I follow what you are talking about. Can you elaborate more concisely?"

Simple, don't mix the content support V10 for those who like the old style V11 for the new style. We need to make V11 docs and tutorials and make sure people can't mix the two. Animals of a different breed if you will…

I am currently setting up equipment for the tutorial task to contribute what I have learned as I now have the room to do so now that the music load is currently calming down.

Master Rat said

Simple, don't mix the content support V10 for those who like the old style V11 for the new style. We need to make V11 docs and tutorials and make sure people can't mix the two. Animals of a different breed if you will…

Ohh I think I understand now; the punctuation was throwing me off. You're saying:

Don't mix the content. Support v10 for those who like the old style, and v11 for the new style.
 

---

While I agree, it's not something I have the resources to do without community help. It is a big time commitment to re-write all the tutorials for v11. Truthfully, not all of them need re-written because some things are still the same in V11 aside from the UI appearance. So I've been taking more of an approach of "find what is inaccurate and fix it".

Also, v10 is no longer supported except for critical bugs. You can read about that here: Status of Composr v11 - Composr . As much as I want to keep v10 going, I would need people to fund me to do so (or they can make their own developer contributions to v10 instead). So as of right now, I'm not making any effort to maintain v10's tutorials except to keep them on compo.sr and (unless Chris has other plans) keeping compo.sr open a while after the first v11 stable release.

I have no grammar I write in musician slang and off-note terms… of slight abuse.