An app driver website

šŸ’¬ Lounge
21

I absolutely use HPM. It's a great PARTIAL solution, because there are many apps/drivers/packages it does not include.

The bottom line, to me, is that many of us wish Hubitat were in the app/driver/package "business", and Hubitat has no interest in being in that business. It's a stalemate between a great little company and some of its fans. Some of us are not terribly interested in searching around on Github and going through wonky steps to install things, and do so only because we like Hubitat more than we hate doing that. So, when we really need to do something, we wade through things and "git 'er done" (but honestly, we really do hate it).

Personally, I use only Apple gear because I can obtain almost anything I need from their app store, with a reasonably high level of confidence that what I download will "just work" (admittedly not complete, so no need for some reading this to post info about the few exceptions). Many will look down their noses at that, but I've just got too much else to do in life to be bothered with a different approach if I can avoid it.

Now, the really big controversial statement: The github approach is a big reason home automation remains a niche hobby for most. It's just too damn hard to find, install and maintain this stuff. But the hard core geeks who love doing it would rather be held captive by aliens than pay for apps in a captive environment OR pay for upgrades, OR pay for a subscription (for things like support). So, we have the system(s) we have, and probably will have them for years to come. There are a few of us who "sit in the middle", buying and using these devices but wishing it were easier, but not that many. Most on this forum are more technically competent, I suspect, and view github as more of a playground. My hat's off to them, but I'm just not one of those hardcore elites of geekdom.

I am a huge fan of Hubitat. I'm really happy I own my four devices. But I'd do more with them if it was easier. So, there... Rant over for a few months. Peace out.

22

Between this and the Hubigraphs conversation... There are thoughts I have around both topics, but hard to formulate them into anything that adds to what has already been said. I'm also torn between my wants / needs as a consumer and appreciation as a developer of the complexities of implementing and maintaining the features being requested.

Kudos to @jshimota for taking on the experiences of those who have trod this path and adjusting your thoughts / comments on this thread. Too often people dig their heels in and won't change their position, but you have shifted, even slightly, which is something to commend, in my book.

(Cue the but....)

I want things to be easier like everyone else, that's why we automate things. But (here it is) there are times where you need to weigh up the effort / $ required to implement and, importantly, maintain a solution. It is only natural that developers gravitate towards the platforms they are used to, i.e. GitHub, rather than building something new. Also, anything we (HE developers) decide to take on ourselves in terms of a app / driver distribution platform has to be something easily maintainable and transferable to other developer(s) in the future.

From my own experience and interpretation of the comments on this thread, it feels like it is the ability to find apps and drivers that is the main issue. There's so many other comments I would like to make, but most have been made already. My only suggestion @jshimota is that HPM has gained what I would say is good support from developers, yes not everyone has got on board, but if you, Bryan or anyone else can build on that success, then I think that would be the best outcome, rather than trying to build something completely new.

Simon

23

I don't think an app store will help that, you can't force people to use that any more than you could force people to use Package Manager.

Isn't that what the "base" (built-in) packages included with Hubitat do? A known stable set of apps that "just work" and all you have to do is click one button in the Apps tab. They don't include every little thing, but they do cover the needs of many people, and get people started.

3 Likes
24

I think @Madcodger is referring to the ability to install apps / drivers that are not built-in, like my Kumo A/C driver :slight_smile: These can't be installed via the normal method, they require something like HPM to facilitate installing the code.

1 Like
25

Anyone can link to any of my stuff on github, and I do try to put it all in hpm (although not everything is in there as I'm lazy).

That said that's about as far as I'm going to go with it. I do the (comparatively little vs other devs) bit I do for free, and I'm not looking for it to take up even more of my time.

I wouldn't put it on a store or paywalled area, as I'm not interested in making money doing this. That adds an expectation of support that I am simply not interested in doing, in favor of the as-is free model.

4 Likes
26

To be fair, HPM could have obscured the use of GitHub more, making it less obvious that is the platform being used to source apps / drivers. If so, then how would that be different to another system that equally does not call out the platform used to store the apps / drivers offered up? The suggested feature does not say it needs to be a user-pays platform...

27

Sure. Make another option, and when it gets as much critical mass as HPM I would obviously consider using that instead of HPM. But probably not both. To me it is more of an either or if it requires developer involvement.

6 Likes
28

Herein lie the obvious issues when it comes to this recurring conversation.

Hubitat Inc. is a company thatā€™s a small fraction the size of Apple or Google. Theyā€™re not in the App Store business.

Except the built-in apps and drivers that provide ease of installation as @neonturbo pointed out, but thatā€™s necessarily limited by the companyā€™s size, of course.

And they have said they donā€™t want to be, or canā€™t, or whatever.

So lots of people would like for it to be easier to find and install custom code that they didnā€™t themselves write. I certainly fall into that category.

Itā€™s easy to ask for that.

But itā€™s not easy to convince a bunch of people from all corners of the world who do know how to write that cod, that there is one preferred way to store and publicize that code for other usersā€™ consumption, then actually create and host it.

So as others have said, if @jshimota can pull that off, then that would be great.

Edit: although Hubitat doesnā€™t have an App Store, they do clearly have agreements with certain community developers that have decided to monetize their own apps and drivers. Presumably thereā€™s a financial (and effort) arrangement that works for both parties there, as well as the consumers/users.

2 Likes
29

Today's my day - I've been moving my house, my office and my warehouse to my new city - a bit over an hour south (I'm in Portland, OR but wont' be tomorrow!). Point is I'll be offline a few days as I reset/reconfigure.

As usual, I think my first post misled or was misinterpreted by many. That's no doubt on me.

What I was try to convey, and didn't, as it was a haphazardly formed idea was this:
If I could go to a website that had a nice clean front end - which gave an attractive listing of all the apps and drivers that flow via HPM but in an easier to see, easier to sort, easier to review format it would be really useful.
In my vision, it would be similar (just sorta/kinda! don't over think) to the Play Store on the Google. It would NOT be a store in the true sense - NOTHING would be for sale. It would simply be a site to assist in finding what is available.
I'm a 'Drupaler'. I write php/sql, code modules, am wicked with CSS etc.. and design sites and have been for ... a long time. I've got serious servers here, and an automated environment for maintanence of dev to prod work.
I say all this, not to impress, but to describe my point of view. I have a focus on use-abilty and GUI.... it's my thing.

  • Imagine a site that showed 'most popular' apps/drivers.
  • Imagine a site that gave a nice search front end (that lets you search tags, categories etc).
  • Imagine a site that provided a background on the devs and what apps they have.
  • Imagine a site that had instructions/walkthroughs on installation, removal for complex apps.
  • Imagine a site that had examples of usage for specific apps.
  • Imagine a site that had a 'gallery' ... a 'show off your dashboard' sort of thing.
  • Imagine a site that allowed an Icon, as well as even an Image representation (developers option).
  • Any link on the site would send you to either back to the github of the app, or, if user has created a profile, to your HPM within your hub.
  • Links to readme's, links to licenses, all would be easier, clearer and quick to find.

I'm not envisioning promotion - simply user assistance and enhancement.
Drupal (or wordpress or... ) all have fairly simply methods to pull data right off what already exists... there is no impact (unless they wish to improve!) on any developer. Yes, I can imagine that dman might want or need to enhance for future adds, but not out of the gate.

Hosting and location are negligible costs. Pardon my vanity but I could literally have a base functionality in a few evenings.

Now, back to the Why. I recently got my GF involved in HE. She sells smart products (blinds) but knew nothing about automation. This led to a completely NON-DEV point of view and feedback. The kind I munge and live on. When I saw through her eyes, I saw the gap between real world users and devs/DIY'ers who 'talk tech'. Is she capable of putting HPM in her hub? Would she know how to find an App? Would she read Forum threads?
All of those are answered " No ".
This community forum is just that, a forum. If you release an App... it becomes lost after a while... unless someone actually comes looking. It isn't Fun, it isn't Easy. It's full of mis-information, old information, and complicated things that help Dev's. In no way is my vision a replacement or change (not that I could, after all it's managed by HE).

Finally, as I mentioned previously, the 'Should I?' question.
When I first posted, I was poking the bear to get feedback, and I got it. RIght out of the block I learned a) others had been before me b) is there really a need or am I just addressing my own wants and desires.

As @sburke781 mention, I adjusted my thinking. I'm now more on the side of 'doable but not necessary'. Who does it help? As Brian pointed out, he built it, no one came. But I do keep in mind that times are changing. I wonder if HE stays and grows, would this fill a need that helps my hobby, and would it be of value. If there's no value, theres no point. And if negative feedback from dev's (rightly or not) taints a project, I'll end up disenchanted and move ala Cobra and others.

So. for now. i'm just thinking about it. Next week, when I've got my servers and workstations back online, I'll probably take a stab at it locally and MAYBE show to a few to garner their thoughts. I guess I'm saying... I plan on playing with this further to see how it fleshes out as a way to better understand how this might work out long term.

thanks to all you fine folk for your feedback!

3 Likes
30

Assuming that "doable but not necessary" becomes "done", here's some thoughts that crossed my mind:

The biggest drawback of any of these is aging... and it's more or less the same at either end. Hobbyists release hobby code without adding them to HPM, either because they don't want a lot of people to see it or don't want the burden of "support."

Not all Developers use HPM... unaware or intentional, the result is the same. Cobra's large collection is an example of intentional. "Scraping" HPM will make an amazing start but how extensible it is will define its uptake, I suspect.

For both groups, when the code goes stale, there's not an easy way to tell that it's not being developed. Thus old/obsolete code can potentially "poison" the site.

I've created a space on the HubitatCommunity github for "abandoned but useful" code and have added one package including adding it to HPM. So the community does have a mechanism to gather the good stuff. The intent is strictly for abandoned but useful in the sense that the code appears complete. No new requests are pending or in demand. If there were, someone in the community would fork it and add it into HPM. However, a minor flaw in HPM is deleting someone else's manifest if they abandon the community.

Also, there's code that won't "fit" into HPM, with Bundles being the #1 example. Bundles are new and not widely used, but they will never get added to HPM.

I "scrape" HPM's manifests hourly, creating a mySQL db each time, and so am fully convinced of the do-ability. But there's a lot of work there for not much benefit. As has been pointed out, the proposed tool would find its greatest use for the newest buyers, which this community doesn't have access to and for which finding the tool will be a treasure hunt without the first clue.

1 Like
31

@csteele Well - from a user point of view, I'd address that with 3 control points. First, I'd track initial add. This will give an 'age'.
I'd track updates. Either by parsing the readme/headers of the code or by release date changes... not sure yet.
I'd add rating system. Commonly, you see the general but well used '5star' or something along those lines. as items decline in value, the rating nearly always reflects this.

Regarding non-HPM code... thats easy as well. I'd give edit ability to devs / users who wish to post info about their apps that are external. May be some moderating there - unclear and not first version but easily added. That would allow for 'external' code reference as well.

Very good info and thoughts! As a known strong entity around here, I appreciate it.

32

:wink:

3 Likes
33

Now you got my attention.

One thing that would have to be mandatory, for anyone using HPM that doesn't want to be part of the hypothetical new site. I do this as a hobby, for myself ... if others find it useful great but I'm not about to subject myself or my code to a ratings system.

Just my 2 cents.

4 Likes
34

To which one of our most prolific developers responded:

So @jshimota, do you see why no one has implemented this before? You should still try, but realize that many of our community developers do this as a hobby - they're not in this to make money (although donations are welcome).

If it were me, I think I would encourage your girlfriend to use HPM and @bptworld's Package Explorer tool. And/or to visit this forum and use the search function, which works remarkably well.

4 Likes
35

I never even knew that exsisted. I think the OP (and apparently your) intent is good, and I would certainly use it. I actually really like the idea. I just don't know how to get developer buy in. Isn't the HPM explorer one of your apps as well?

1 Like
36

If we had it originally will this be an update? Or do I need to delete and reinstall?

37

Same app, just do an update.

Thanks

2 Likes
38

Well now, there you go! Killed it before it got out the gate! Saved me some effort for sure. No discussion, no creation. Just a foot stomp. Love it. Dev's - it's all good. I knew it was a dark hole the moment everyone went sideways with the intent.

As for the GF learning HE... ah. no. as one of our most prolific developers responded: "I'm not about to subject myself ...' HA!

(and I just got done studying the Yaml/json import content. would have been great fun - sorta like my battery app on google!)

FOr the next idiot like me that peeks down this hole - here's the import function all prepared:

39

I'm not so sure it should be written off so quickly. Ideas have been exchanged and questions presented. Doesn't mean it's not doable, just may take some more thought to keep things current or relevant.

I can see why "hobbyist" devs don't want to be subjected to a rating system. Get a few people that don't understand how the apps work,or don't take the time to work with the dev or community to figure it out and it won't be long before a good dev gets completly trashed for no reason. Think about the times you have seen a one or two star rating. How often is it really obvious that who ever wrote it had no clue what they were doing and trashed the subject as a result. Maybe using update dates may be an option? Just spit balling. It's an intriguing idea, worthy of discussing further.

3 Likes
40

I donā€™t see the relevance. What does a Hubitat developer not wishing to participate in a rating system have to do with your girlfriend learning to use her Hubitat hub?

Thereā€™s plenty of 5-star products on Amazon boosted by paid-for reviews. Conversely, Iā€™ve purchased unrated products that have worked out just fine.

2 Likes