Announcement

Collapse
No announcement yet.

FEEDBACK: What is holding you back from HS4?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    FEEDBACK: What is holding you back from HS4?

    Hey Devs,

    This topic has the intention of gathering feedback from your experiences and your ideas.

    Aside from time, are there any specific things holding you back from updating your plugins to HS4. I know there have been posts concerning bugs or missing features. If that is still the case, share those specific bugs or features here.

    We are looking for feedback on the SDK to improve and expand. I know some of this feedback is scattered throughout the forum but I'd like to pin it to this topic. Lets keep it constructive and forward-thinking. Please let us know. We are looking forward to your feedback.

    ~Jon

    #2
    1/ Speakproxy support. Believe you were going to simplify the interface which is currently the HSTOUCH.SpeakerClient
    2/ Player Device tile view
    3/ I'm still struggling with creating html pages, especially doing AJAX updates on sections which are created with scriban enumerations and liquid tags. How do you re-generate a section like to that have it dynamically updated
    4/ Do we have any idea what is to become from HSTouch? Is the mediaAPI completely dead or are you guys thinking about something else allowing your users to create great mediacontrol screens?
    5/ how do we migrate our users from HS3 to HS4, especially when controls on root devices need to turn into features etc. I'm also not clear on how events created in HS3 migrate to how it is suggested to be used in HS4

    Comment


      #3
      In order of importance:
      - Documentation
      - Bugs that are open for months
      - Not enough/no support here on the developer forums :-(
      - Revenue vs time
      - Missing (advanced) examples

      Regarding bugs:
      - still no way to point out the main feature
      - extradata information is somewhere changed to lowercase (link)
      - TODO's
      - EDeviceType that appears to do nothing (link)
      - ColorPicker that is still not working

      Some of these have been promised for months ...

      Comment


        #4
        Originally posted by kriz83 View Post
        In order of importance:
        - Bugs that are open for months
        Are those in github? What bugs that have been open for months are you referring to? Are there specific ones that are holding you back or is it just that there are bugs? Please be as specific as possible so we can prioritize and act on these issues holding you back.

        ~Jon

        Comment


          #5
          Documentation for device creation. A LOT of time was spent on trial/error.
          An immense amount of code changes going from clsPageBuilder (HS3) to the new bootstrap stuff for the sake of mobile-ready.

          Also, a plugin really doesnt need to be mobile-ready - does it? If you are configuring things and the plugin just runs, why make it mobile-ready?

          Documentation in general sucks as well. The NetCam plugin and the sample plugin used 2 different Ajax post examples further complicating the learning curve.
          HS4Pro on a Raspberry Pi4
          54 Z-Wave Nodes / 21 Zigbee Devices / 108 Events / 767 Devices
          Plugins: Z-Wave / Zigbee Plus / EasyTrigger / AK Weather / OMNI

          HSTouch Clients: 1 Android

          Comment


            #6
            Jon,

            I have several items in github. Only one day in this year issues have been responded to , including at least one issue where I think there should be more attention for. Some even got closed without an chance to explain why it would have been an issue for the PluginSDK.

            My highest priorties on github are these, with emphasis on "my": Others sure will want attention on their issues.

            Very important for my conversions: https://github.com/HomeSeer/Plugin-SDK/issues/75
            Most important question: https://github.com/HomeSeer/Plugin-SDK/issues/71
            Most annoying: https://github.com/HomeSeer/Plugin-SDK/issues/68
            Bothering existing users with existing events: https://github.com/HomeSeer/Plugin-SDK/issues/72

            Have been trying to push these several times, but no response for a long time. I saw a statement passing by last week that someone would look at it. But no response till now.

            Wim
            -- Wim

            Plugins: JowiHue, RFXCOM, Sonos4, Jon00's Perfmon and Network monitor, EasyTrigger, Pushover 3P, rnbWeather, BLBackup, AK SmartDevice, Pushover, PHLocation, Zwave, GCalseer, SDJ-Health, Device History, BLGData

            1210 devices/features ---- 392 events ----- 40 scripts

            Comment


              #7
              For me unfortunately it was the lack of real interaction between HST and developers before the changes came in (combined with the fact I don't see much in the way of point upgrading my own HS install to HS4) - such as the loss of multi instance plugins, changes to the UI, changes device structure etc. If I had been warned some time before then perhaps I would've made gradual changes to my HS3 plugins as they had periodic fixes. They all appeared to happen without much in the way of warning, for me the changes required to make my plugins HS4 native are considerable and at the minute unjustifiable.

              If the changes needed were going to take me a couple of hours per plugin (best estimate is they will take me days, I'm not that good and have to make sense of the documents to start with) then perhaps I would consider it, as it stands it is difficult to stomach that effort. I appreciate you said that you didn't wish to hear about time being a factor but in reality time will be a significant barrier to many when combined with other factors.

              Also I have no doubt the new abstract layer is with the best of intentions but to me it seems to have made it more complex rather than less so, the HS3 SDK gave me initial difficulties that was more down to the examples. Once I got my head around HS3 it I understood what was going on and it was in reality quite simple, now I have to learn a new trigger/action interface for what appears to be primarily a UI change in HS4 compared to HS3 - I'm not then clear what I am gaining.

              One thing that might make it easier is hosting another webinar, Rick/Rich/Wade did a HS2-HS3 webinar and I found it quite valuable. It was two hours and covered most of the things and people could ask questions. Two hours time from the developers would probably answer many hours of third party developers time.

              Comment


                #8
                TBH Jon, currently I am in a phase that I refuse to put any more effort into this. GitHub is dead, documentation is zero, code has lots of TODO's and threads in this part of the forum often don't get any response from HST.

                I agree with mrhappy that I do not see the benefit of updating my plugin for only some UI changes, which for the most part are not beneficial at all.

                IMHO, HST should have done their homework much better. Way too many (obvious) issues are found by users. The NEVER action got deleted, device linking is gone... How is this possible? How can these things not be noted by your testers? Next to that, i really do not like the new interface.

                Sorry for the ranting, but as a long time HS user I am very annoyed by the lack of professional software development that has come with HS4.

                Verstuurd vanaf mijn SM-G965F met Tapatalk


                Comment


                  #9
                  Documentation, Documentation, Documentation.

                  I need more written words in the SDK:
                  - code fragments and explanations in the SDK for each property and method
                  - for features that were depreciated (ie multiple instances, and other basic properties like devicestring) we need recommendations for how to do them in HS4
                  - what do we need to do for hs3 users moving to HS4. we don't have dummy hs3 root devices in all cases.

                  since my hs3 plugins are hs4 compatible, there isn't the urgency like we had from hs2 -> hs3

                  Mark

                  HS3 Pro 4.2.19.5
                  Hardware: Insteon Serial PLM | AD2USB for Vista Alarm | HAI Omnistat2 | 1-Wire HA7E | RFXrec433 | Dahua Cameras | LiftMaster Internet Gateway | Tuya Smart Plugs
                  Plugins: Insteon (mine) | Vista Alarm (mine) | Omnistat 3 | Ultra1Wire3 | RFXCOM | HS MyQ | BLRadar | BLDenon | Tuya | Jon00 Charting | Jon00 Links
                  Platform: Windows Server 2022 Standard, i5-12600K/3.7GHz/10 core, 16GB RAM, 500GB SSD

                  Comment


                    #10
                    The last HS3 Release 3.0.0.548 was made on 8-12-19. While HST has not been advancing HS3, there has been much activity with 3rd party development for HS3 since that time. Once an HS4 version of a plugin is made then the developer has doubled their workload because updates are needed to both the HS3 (active) and HS4 (future) versions of the plugin. Bouncing back and forth between the two APIs is not easy. Relearning is needed each time, but at least the iterations have the benefit of having a reference of having done something else similar in the plugin.

                    My plugins are typically characterized with significant configuration options with the end objective of providing the user a HS device for deployment of what was setup. The day-to-day use of the HS device may be fine for a mobile user. This same user will give up on the setup process with their smartphone. Setup layouts that have been designed for large real-estate screens will be total frustration for the mobile user. Making it worse is the mobile views on the desktops take up more screen space so what would fit nicely without scroll bars in HS3 plugin now needs scrolling in its HS4 version.

                    The HS3 version of my plugins will provide an easier setup experience on the desktop and the Device views presented by HS4 will be the same if setup via HS3 plugin or HS4 plugin. There really is no motivation to rewrite any more plugins until HS4 is the dominant version of HS in the field and even then only after there is a breaking change that prevents continued use of legacy plugins. Hubitat was a relatively easy conversion since it has a very minimal UI. mcsMQTT has been a real bear and as stated above doing everything twice on this very active plugin is a downer.

                    I have recorded all the deficiencies that I have uncovered in Github so no point in repeating them here to add yet another tangent thread on the same subject. The pointers to these are:
                    https://github.com/HomeSeer/Plugin-SDK/issues/51
                    https://github.com/HomeSeer/Plugin-SDK/issues/77
                    https://github.com/HomeSeer/Plugin-SDK/issues/79
                    While it is not holding up plugin release I find https://github.com/HomeSeer/Plugin-SDK/issues/70 a major usability issue.

                    Comment


                      #11
                      My biggest hurdles (besides time) are:
                      • Dropping support of multiple-instance plug-ins. I will have to rewrite quite a bit of code to:
                        • Take into account for which device a request is.
                        • Change the GUI to allow management of multiple devices.
                      • Adaptation to the new device-with-features model. The devices I wrote a plug-in for do not really have a primary feature. Sure, I can move the root status to a feature and make that the primary, but the root status is the least interesting information of that device. You only need it when troubleshooting and not during normal interaction with the device.
                      • The struggle that other developers are having to get basic things working using the HS4 way.
                      So I have not yet published my plug-in for HS3, because any additional effort in the HS3 version would be wasted. And I think I won't port it to HS4 and publish it this year. HS3 and the plug-in in its current state are working for me, and I will probably only put more effort into it if HS3 can no longer be used for my (simple) wishes.

                      Comment


                        #12
                        Obviously we are being pulled in many directions at the moment, but let's move forward. Rather than spending a week writing docs that only address a portion of your issues, please post what you need in order to finish your port, or work on a new plugin and we will do our best to assist. You can also PM me directly if I don't get an answer here in a timely manner.

                        I can share a little info with you that might help. I know some don't think many users woud upgrade just for the mobile changes. Right now over HALF of our user base has purchased an upgrade to HS4. So no doubt the interest is there!

                        In regards to HSTouch. The current version works with HS4 just as it did with HS3. We are looking at creating a new version with a web based Designer and new clients. So at this time we do not expect to make any changes to the current version.
                        💁‍♂️ Support & Customer Service 🙋‍♂️ Sales Questions 🛒 Shop HomeSeer Products

                        Comment


                          #13
                          Jon, I understand that currently it's all hands on deck, but this is largely due to the choices made by HST. Documentation (and testing - I keep on repeating it) should have been done from the beginning. There is no excuse for this. Also, we have been promised documentation more than once ... A good SDK comes with even better documentation. It is so easy to create good documentation practices in Visual Studio, it's a shame they are not used. Within our company, documentation and tests are mandatory. Got no documentation, then your pull request is rejected. No tests created/modified/changed ? We don't even start the review then ! Simple, and it makes sure we adhere to a certain quality level (static analysis, stress tests, ...)

                          Same with Home Assistant. If I deliver changes to their archive, they are rejected if there are no tests attached to them. On the entire SDK GitHub page, there is no test to be found. To me (as a professional software developer), that explains why we have a huge amount of issues :-(

                          I was really enthousiastic when the SDK was published on GitHub, but when you publish something on GitHub, you have to make sure you maintain it. There are over 30 open issues, and no commits for 8 weeks. The way I see it, that is some very bad publicity.

                          Originally posted by rjh View Post
                          I can share a little info with you that might help. I know some don't think many users woud upgrade just for the mobile changes. Right now over HALF of our user base has purchased an upgrade to HS4. So no doubt the interest is there!
                          That doesn't say much. You had a very good discount price, so I can imagine a lot of people pulled the trigger. I'm interested in new sales, not the updates. But even then, my plugin should just work on HS4, I am really not interested in spending hours (if not days) for a bit of extra beer money.

                          Comment


                            #14
                            Originally posted by rjh View Post
                            Obviously we are being pulled in many directions at the moment, but let's move forward. Rather than spending a week writing docs that only address a portion of your issues, please post what you need in order to finish your port, or work on a new plugin and we will do our best to assist. You can also PM me directly if I don't get an answer here in a timely manner.
                            .
                            Rich,

                            I have posted my major github issues in post #6, that is including answering questions on the github.
                            These should be combined with the open issues on the plugin submittal page:
                            • production HS4 plugin version 4.0.0.4 submitting 4.0.0.10 gets complains that beta version is lower then production version (known by Jon)
                            • Production HS4 store picture, features and description are not published in the store (known by Jon)
                            • Plugin installation on existing HS3 license is not asking to purchase upgrade (known by Rich)
                            If these are solved I guess I am pretty much complete, except for the strong wish to have a complete documentation on how to make a new plugin

                            Wim
                            -- Wim

                            Plugins: JowiHue, RFXCOM, Sonos4, Jon00's Perfmon and Network monitor, EasyTrigger, Pushover 3P, rnbWeather, BLBackup, AK SmartDevice, Pushover, PHLocation, Zwave, GCalseer, SDJ-Health, Device History, BLGData

                            1210 devices/features ---- 392 events ----- 40 scripts

                            Comment


                              #15
                              We can complain or we can move forward. I cannot change the past, I can only offer to help you moving forward. What can I do to help you get your plugin completed/ported? Which Github issue is holding you up, we can look at those first.
                              💁‍♂️ Support & Customer Service 🙋‍♂️ Sales Questions 🛒 Shop HomeSeer Products

                              Comment

                              Working...
                              X