Future of OBEX
VonSzarvas
Posts: 3,391
Hi all!
I'm wondering what the ongoing interest in OBEX is ?
The annual server rental and software plugin renewals are coming up very soon, and whilst considering the costs I'm thinking about options to keep it funded, and whether it's even wanted at all?
Edit: Parallax have made a donation toward the running costs for another year which was enough to keep it operating. If anyone needs help with OBEX, contact support@parallax.com.
Comments
I wish you never got rid of the original Obex
I agree with DigitalBob. Von you need to consider that the "how to examples" are your bread and butter... it you take that away then it just a matter of time before other things higher on the ladder will also fall away. I say keep the OBEX and if anything promote it as best you can. ... just my Nickel's worth
I'm rather new to this community, arriving in earnest during the git repo period - which worked well for me, but I understand why it may have been cumbersome for those who prefer zip files. I find the current OBEX quite effective at searching for pertinent options based on part number, peripherals and language. It seemed like there was a lot of effort to develop features of the current OBEX to satisfy user needs. I can't fathom what was so special about the original OBEX.
Can someone enlighten me?
Also, how much does the OBEX site cost?
How hard does it get hit?
How big is the database/storage?
Finally, having some public place to share code is absolutely essential in this community, especially the high quality essentials contributed by the generous members.
There certainly needs to be a public place to share code for Parallax products. The current OBEX is better tailored to this community than generic repositories like GIT. The current OBEX has only been viable/useable for a relatively short time. Give it at least another year before considering chloroforming it.
I agree too. A head banging moment in history.
$300-$500 a year (depending on which plugins need renewing, and which can just sit at old versions), 3000 or so unique visits per month, storage ~20GB.
Well, not my bread and butter - far from it sadly (I'm merely a volunteer here). But I agree on the point.
The original concept for the new OBEX site was to have an active growing repository for all things Propeller, with weekly updates to include: How-to's / demos / drivers / mini-projects & examples for building projects and integrating Propeller into customer designs, PCB layout sample files, and all things that an engineer or new user would need to lookup to get started and to work with the Propeller. Alongside that, a searchable docs system, initially intended to include the P1 & P2 "API" docs, and then regularly expanded each week, along with the other content.
Whilst rugs have been moving and work never really started on the content, I remain hopeful that at some point things will change for the good and that the OBEX strategy will form part of the future growth for Propeller.
So, I put old and new here and fail to see what made the old one so much greater. or was there some political turmoil around the recreation that underlies the comments? No matter, at this point. I'm just trying to understand the issue and think about what can be done to avoid losing access to all the code up in OBEX now. Are you just asking us to pass the hat to defray the costs?
In either case, the extra bits seem to be getting rolled into Quick Bites and the improved collection of links on the Documentation page. While not live community driven, are a welcome improvement as it evolves.
@VonSzarvas are you paying this hosting cost? Did you build the new OBEX site? (on what platform? it seems quite expensive to me for light resource demands of 100 users per day and ~20GB) Has it been difficult to satisfy particular desires of this community for what they want in a community code share system?
I'd put it all in a Fossil source code management system - functional, serviceable, if not the perfect bespoke system and a bit old-school in interface and management. This would host easily on a $12/month Digital Ocean droplet with no extra charges.
I'm not trying to challenge anyone here and all the effort that has gone before. I truly appreciate what we do have available. Just trying understand where the issue is and how I might contribute.
it seems unlikely, or un-necessary to wrest the community activity away form the only lively spot - on the forums. too bad that is such a difficult spot to find current best hosted files.
no-comment, but, yeah it seems something like that....
Last year I paid 50%, Parallax 50%. This year I'm finding it harder to contribute for sure, and I've yet to find out if Parallax are ready to fund any part of it. Hence the research to gauge the community interest in the thing continuing. I'm still trying to wrap my head around if having a community funded resource would be practical. Ultimately it's all a waste of time anyway, if there's no ongoing content development and membership management going on; the tool should be leveraged for marketing and growth of the Propeller, which is more a commercial aspiration (or should be).
Yes I did- it's on Wordpress. The server was sized for larger load and features, and possibly could be reduced now- but that also takes time and attention. And operationally there's no one watching and resourced to react when it overloads/etc, so having it slightly bigger takes away those concerns without adding much cost to hosting, and also keeps it snappy. Anyway, the server is ~$20 a month- then the domain name, mailsender, storage and software renewals round out the total price. If I do end up paying for it again, then I probably will drop it down a level or two and just let it run slower. But I'm hoping that won't be needed, as I like it snappy
Was moderately difficult to manage it with the given requirements,- like everything free, no custom code and only off-the-shelf plugins, and then- with lot's of requests for custom behavior, etc..
But we kinda got to a compromise with what it is.
Yeah, me too. And I appreciate you also thinking about it. First up I think I need to chase Parallax for a decision, and then go from there. My feeling that it's an important resource has been well supported by everyone in this thread, and I'm glad for that.
Digital Ocean offers some nice packages ... I have a several running at the moment for $6 a month and you don't need a 3rd party host like GoDaddy or HostGator etc.. What you do need to be able to do is build a server environment from the ground up (My preference is Apache) Once you have that then you can either hard code the HTML yourself (my preference) or use something like Wordpress. As far as a domain, you attach that to the Digital ocean IP just like you would any other domain name. The Digital Ocean route would certainly be cheaper, but may not be the right fit for you as you would need to be familiar with Perl,Python, HTML, JavaScript and a few others to make it all come together. If I had the time I would offer to do it as a contractor. I'm sure there are several folks here in the forum that would be willing to work on this but if I were them I would only do it from a Contractual point of view. There is too much work involved to be a courtesy project.
I am however getting vibe that you are being ghosted from Parallax ... what's up with that?
Parallax has their limited hands full with recovering from ransom ware attack last week. So, I'd go easy on them.
I have numerous $6~$12 / month digital ocean droplets for home and consulting. I tend to stand up Fossil instances (https://fossil-scm.org/) handling user management and communications and code and ticket handling behind Caddy server (https://caddyserver.com/) dead simple to install and handles the TLS cert stuff automatically. I use Fossil to manage and document projects for clients and myself. I have many many of them setup and I've load tested and can confirm the claims of performance of that scenario. not doubt it could handle the load of all of the Parallax community on a $12/month droplet. I've seem some heavily reconfigured instances of Fossil that you can't even tell it isn't a hand crafted site. Though it'll never be, and tries very deliberately to remain a fast simple (little to no javascript) service for low-end browsers. having said that, I've also served a static blob of a javascript application with embedded data out of a Fossil. Ultimately, Fossil is intended for work around single projects, and would function pretty much the way the github repository attempt did.. with more/different community features. Anyway, that's just me. 'can't tolerate the heavy frameworks and associated administration of having to adjust to the whims of the framework builders.
Current OBEX is snappy and attractive. but I do understand the effort to build out all the other community features is a major undertaking, let alone the community management and moderation. There is quite a variety of developer vintages in the Parallax world, and we're set in our ways, and rather be doing work than figuring out some new system to get around in. I sure don't know how anyone would receive a different new system. I'm curious to noodle around with a few OBEX community ideas though and see where it leads. Is there a way I could get a data dump of the projects/attributes on the OBEX site? if it is a hassle, I could just scrape a few dozen packages to get enough to play with various schemes.
I just know we need to keep something up there. Do you have a Patreon, or something like that? That's how I'm contributing to other key Parallax developers.
Having the system in Wordpress was a specific requirement Parallax made, so changing to another backend might not be a solution here. Also my instinct would be that potential savings of $8 a month ($96 a year), would easily get eaten by the labor cost of changing systems, etc... (We probably already did that time just talking about it
I'd rather put any free time into adding to the content we have, rather than changing what we already have.
I guess the best help right now would be to find someone (or a group) willing to build out the docs content on the OBEX portal for free? Or to build out the example projects and how-to content for free?
I concur all around.
I've seen effort of contributors recently to update their OBEX content. The documentation pages seem to get steady improvements without much fanfare. I continue to find them useful to revisit instead of just re-viewing the, potentially stale, docs that I had already downloaded. Otherwise, though it took me a little time to warm to it, the Forum seems to be where the action is for working on problems and showing off projects.
The situation is clear to me now. Sorry to have taken so much of your time on this.
referring back to your original question: Yes, it is wanted/needed, it needs funding.
Certainly, Parallax understands that.
What's wrong with a Dropbox or such like?
Craig
That's It around $300 a year, that's less than $1 a day. I spend more than that for coffee per day. I'm considering sending you the $300 for the Obex.
If people are supposed to do something for free, it is essential, that Parallax states commitment to keep the work accessible!!! Writing down documentation for something is a lot of work! There are many many hours of work from many contributors in the Forum and in OBEX. Unfortunately Parallax seems not to see how important documentation, examples and libraries are, because they enable customers to build their project very much faster.
On the other hand I understand, that Parallax has to have the cost in a reasonable relation to their revenue from the Propeller business. If the revenue is so low, that they cannot afford this type of OBEX (it's nice!), then it will (sadly) be better to switch over to something, that can be afforded for the next years.
Ah, there might be misunderstanding of which documentation is being discussed.
I would happily participate in some kind of crowdfunded (Patreon?) system to keep and expand OBEX. I wouldn't even care if it remained full access for free for anyone. I suspect I am not alone in this. IMHO OBEX should be a one-stop shop for objects, docs, QuickBytes, other vetted cool stuff. Heck, why isn't there some cool stamp stuff there. Where do I send the money!
Update: The feedback and enthusiasm from everyone for OBEX was more than enough to convince Parallax to keep it going !
This doesn't answer to how we might keep the content expansion moving forwards, but after the summer break that is something to loop back to. Jeff Martin is quietly working on building out the P2 docs in the background, and Stephen & Chip are working on the new compiler; all those things will soon come together and provide more new material.
We keep on rocking !
Now, that's good news! I had considered posting on the state of the P2 documentation but had resisted, knowing how busy Parallax was with other things.
Indeed you do
Cheers to everyone's continued work ... but GO JEFF!
Hoorah! We have Ken onboard for another year.
That makes it worth me (and hopefully everyone who can help) continuing to improve the content. Thank you Parallax!
https://www.parallax.com/new-object-exchange-obex-for-propeller-1-and-2-multicore-microcontrollers-is-loaded-with-source-code/
When you click on an author name, presumably to see other code from that author, you simply end up back at the main page. Let's take the example (screenshot) posted above. The first author is Jon McPhalen. When you hover over that name, the target link is: https://obex.parallax.com/author/jonnymac/
However, when you click that link, you're taken to: https://obex.parallax.com/
Logging in will unlock more features; is that the situation - that you didn’t login before trying the author links?
I like the new obex. What was the problem with the GitHub version? No way for people to add to it themselves?
I personally liked the GITHUB concept but not the implementation.
Each object should to have a separate download link, as far as I could tell you could download the entire repository but not an individual object.
If the objects were setup so they were searchable (we could search but it gave code as a result and not a readme file) and individually downloadable, then it would work for me. I don't really want the whole thing, just the object/s I want. We could then possibly offer suggestions for code improvements via GITHUB for that individual object and help keep the code up to date.
I know I can download the lot so I have a local copy and GITHUB and can then keep an eye on the original and keep the 2 up to date, but I am less sure how to contribute changes when you do that.
Perhaps it's an education thing as GIT simply wasn't a thing when I started programming.
Not sure how relevant this still is with the new obex, but in order to contribute you either had to:
In both case human interaction is needed on the receiving end.
There is/was a possibility to only download a zipped part of a repo (look at the suggestions here: https://stackoverflow.com/questions/7106012/download-a-single-folder-or-directory-from-a-github-repository. I think I tried one of those once upon a time, but not sure any of them still work)
The obex frontend should be similar to how it is now, where you can download a zip file for each object, but the backend should just be a Git repo. There could be some automated way to allow object owners to merge branches that only change their own objects back to master. That way, people can contribute to or access the obex either way.
I think my comment was relevant because the point of this discussion was asking if the OBEX is still useful and perhaps how to make it cheaper.
My point was that if the Parallax GITHUB was organised like say, the Sparkfun GITHUB site, then each OBEX item would be its own repository.
If you click on one of these you get a description of that item, just like OBEX, it could even be much more detailed than most are at the moment, you could also have a link to the releases, both current and historical if you liked. Then the OBEX could just be an index to the appropriate place on Github, if it was needed at all.
Von,
You might want to update the opening post to say Parallax is now funding the running costs of Obex.
Done. It's funded until 24th July 2025.