Feature point questions, re: Janet

Greetings Phil; I had asked Janet if I should contact you directly with more engineering-type questions and she suggested I contact you via the issue reporting system...OK, seems odd but here I am.

One thing thats sort of vital to my understanding of how this app is designed involves configuration. Not the configuration file itself but rather my rather poor understanding of modern development wisdom. I am sure much of it stems from my 80's/90's/2000s dev mindset. So I am asking. In truth the config file seems to be more like an app template, in the same way a makefile was. You change your config/makefile and us it to rebuild your app. In this perspective, the entire app is hard-coded to only do what was in the original "config" or template file.

Config files as I know them on Linux/Mac/Windows/Sun/DOS/CPM et al were just that: you could reconfigure the app simply by loading a different configuration at runtime, no changes to the binary. That seems 180 degrees from what I am seeing here.

So I am forced to ask just so I am clear on your thinking: is it the intent that each "use" of this will consist of a uniquely-built binary and assets (pics)? I think this is the disconnect from your comments on the use of the QR code reader. the intent was not to use the camera to download something from playstore, it was to embed what you call a config file and assetts into a QR code, the idea being the person has a running version of your app on the phone, sees a QR code promising help with (say) speaking at a doctors office; the user tells the app to import, the app accesses the camera, decodes the QR code extracting the new config and assetts if needed (thinking there might be stock assets that come with the app covering 90% of the common situations; the special case assets might be extracted from the QR code or maybe simply image links that the app can then download and use with the new configuration.

So long use-case made short, the user has your app with very generalized images/texts, decides he or she wishes to go to out to the shoppes for some fish and chips; they import a new config via QR code and instantly, the previously very generalized app now has finely-tuned images and responses unique to that situation. The user gets alot more out of the app because it can be so easily specialized for non-standard situations.

But none of that makes any sense if the app will always be hard-wired to do one specific set of responses. I know I can change them via git push but the user, and my only other tester would quit before going thru all this to reconfigure the app, and I suspect most mere mortals will too.

Thoughts? Clarification? What am I missing. I will be the first to tell you that while I was good at this not long ago, the dementia makes strategic thinking needed for coding and design painful to the point that its impossible....

Cheers Jeff

Hi Jeff:

Each source file.txt creates a new version of the app each time it is updated.

You can have as many sourceFile.txt in a single repository as you wish as long as each one is in a different folder. This lets you have as many variants of the app as you wish as the folder name is appended to the app name to differentiate it. If you give each app a different title and icon they will be easy to recognize on your android screen. So there will be one app for going to the supermarket, another for going to the movies etc.

At the moment each app is limited to one level of display. I will improve on this as soon as some-one starts producing apps that cannot fit on one page.

Each word tile can also have a backing photo named via the http://www.appaapps.com/howToWriteAnApp.html#cmdKeyDef_photo_url to make recognition easier.

I believe our most immediate need is for some-one to actually produce a few apps by themselves and try to use them in the real world so we can get some feedback to see which of the many possible improvements we might make with our very limited resources is in fact the most pressing one in reality.

Please let me know if there are any more questions I can answer for you that would lead us to this important goal?

Thanks,

Phil

Phil; Wow...design must have changed alot since I got out. This feels like having to have a prebuilt binary of (say) VLC, recompiled to play each individual movie, clicking on VLC=Matrix, VLC=Avatar, VLC=Avengers, etc. A lifetime ago I had a user, unfortunately for me also a relation, who called requesting emergency assistance because her 20 meg (RLL no less) drive ran out of space and all she had on there was Word Perfect and 20-30 documents. She swore that was all that was on there yet she was out of drive space and could make no more docs. Upon examination of the drive I find that every time she wanted to create a new document, she was creating a new folder for that new document, and then installing an entire copy of Word Perfect (of the time) in that folder, one install per each 30K document she created! Facepalm anyone?

So to my decrepit brain this strategy seems really like a hard way to go, recompiling for each use-case...but then, when I was big in dev, knowing solid OOP stuff made me seem weird to the old guard of monolithic development, some of these guys never made the leap and remained in the land of mainframes and COBOL until they died. So for all I know, I might seem like the chimpanzee, throwing rocks at the moon to any current developer.

I will tell you one other thing for our interactions. Dementia means my brain is breaking down. I am enough of a systems troubleshooter to recognise the fact my brain may not always give me good data, or think things through clearly. I accept that on a pretty pragmatic level. So if someone states something and I think its off, I may say something but never make an issue of it simply because I know my brain is faulty and if I state my case on some subject and it appears wrong to those about me, I accept it as another mental failure and just move on, accepting what to me seems to be an absurdity as a new facet of my life. Thats not giving up, thats me preparing for the coming hallucinations and delusions common to dementia. If I can discard what seems odd to me now, it should be that much easier when I am trying to convince myself the dragon in the room is just an illusion and to ignore it. Like brain-damaged presidents, eventually they all go away.

So. In order to not just support what I just such but to hopefully aid your project, I will build some simple apps for sort of proof of concept testing. One caveat is, until this is capable of exporting the output to each channels (IOW sharing to email, text, even voice) and until it is capable of ...chaining together or stacking responses, only the most fundamental phases will be used. For something to work in a work-place environment, the responses might need to stacking or caching....(note, my aphasia makes this hard so allow me to be clear: I am meaning where the user can select and cache multiple responses which are then played back as a single response. Very simple example, say a disabled worker needs to inform his or her boss that they will be gone to visit the doctor the next day but will be into work in the early afternoon. That statement is too precise/specific to be compiled into anything but the elements of it (I will be gone, (date/time), (reason), I will be in after lunch) are 5 or 6 individual responses, making no sense when uttered on their own but when cached and then spoken as a single response, sounds like a more human replying than a soundboard.

OK I shall put some effort into working out some specific use-cases, sadly they will be more US than UK-centric but its what I have to work with and the point afterall is to give you some ideas about the most useful applications of your system. I have done a pull (get, whatever the current term is) and will work on it this afternoon. I raise cannabis for my condition and as such, my morning hours are usually consumed with that.

Thank you so very much for listening. I was a coder for a long time and you know that means thats who you hung out with...but now that I can't code I am stuck with mere mortals and so have very much relished our conversation. Like a breath of fresh air for me; you don't realise how much you miss it until you are no longer around your peers and have to be content debating the cost of petrol or what form of dementia my elected officials have. All subjects that get tired after a single sentence. Design? all day long. Smile. So thanks. So you know I am going to try to find a way to build the app to talk to another hacker....could be interesting...

Excellent! Once we have the most basic app working we can start to make incremental improvements to the system to meet real world requirements. Please post questions that occur to you when trying to build an app to GitHub so that we can record the process for posterity. You might find: http://www.coreliu.org/howToWriteAnApp.html#sft helpful.

On Thu, Jun 14, 2018 at 4:45 PM, Jeff notifications@github.com wrote:

Phil; Wow...design must have changed alot since I got out. This feels like having to have a prebuilt binary of (say) VLC, recompiled to play each individual movie, clicking on VLC=Matrix, VLC=Avatar, VLC=Avengers, etc. A lifetime ago I had a user, unfortunately for me also a relation, who called requesting emergency assistance because her 20 meg (RLL no less) drive ran out of space and all she had on there was Word Perfect and 20-30 documents. She swore that was all that was on there yet she was out of drive space and could make no more docs. Upon examination of the drive I find that every time she wanted to create a new document, she was creating a new folder for that new document, and then installing an entire copy of Word Perfect (of the time) in that folder, one install per each 30K document she created! Facepalm anyone?

So to my decrepit brain this strategy seems really like a hard way to go, recompiling for each use-case...but then, when I was big in dev, knowing solid OOP stuff made me seem weird to the old guard of monolithic development, some of these guys never made the leap and remained in the land of mainframes and COBOL until they died. So for all I know, I might seem like the chimpanzee, throwing rocks at the moon to any current developer.

I will tell you one other thing for our interactions. Dementia means my brain is breaking down. I am enough of a systems troubleshooter to recognise the fact my brain may not always give me good data, or think things through clearly. I accept that on a pretty pragmatic level. So if someone states something and I think its off, I may say something but never make an issue of it simply because I know my brain is faulty and if I state my case on some subject and it appears wrong to those about me, I accept it as another mental failure and just move on, accepting what to me seems to be an absurdity as a new facet of my life. Thats not giving up, thats me preparing for the coming hallucinations and delusions common to dementia. If I can discard what seems odd to me now, it should be that much easier when I am trying to convince myself the dragon in the room is just an illusion and to ignore it. Like brain-damaged presidents, eventually they all go away.

So. In order to not just support what I just such but to hopefully aid your project, I will build some simple apps for sort of proof of concept testing. One caveat is, until this is capable of exporting the output to each channels (IOW sharing to email, text, even voice) and until it is capable of ...chaining together or stacking responses, only the most fundamental phases will be used. For something to work in a work-place environment, the responses might need to stacking or caching....(note, my aphasia makes this hard so allow me to be clear: I am meaning where the user can select and cache multiple responses which are then played back as a single response. Very simple example, say a disabled worker needs to inform his or her boss that they will be gone to visit the doctor the next day but will be into work in the early afternoon. That statement is too precise/specific to be compiled into anything but the elements of it (I will be gone, (date/time), (reason), I will be in after lunch) are 5 or 6 individual responses, making no sense when uttered on their own but when cached and then spoken as a single response, sounds like a more human replying than a soundboard.

OK I shall put some effort into working out some specific use-cases, sadly they will be more US than UK-centric but its what I have to work with and the point afterall is to give you some ideas about the most useful applications of your system. I have done a pull (get, whatever the current term is) and will work on it this afternoon. I raise cannabis for my condition and as such, my morning hours are usually consumed with that.

Thank you so very much for listening. I was a coder for a long time and you know that means thats who you hung out with...but now that I can't code I am stuck with mere mortals and so have very much relished our conversation. Like a breath of fresh air for me; you don't realise how much you miss it until you are no longer around your peers and have to be content debating the cost of petrol or what form of dementia my elected officials have. All subjects that get tired after a single sentence. Design? all day long. Smile. So thanks. So you know I am going to try to find a way to build the app to talk to another hacker....could be interesting...

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/coeusprime/speech/issues/12#issuecomment-397342951, or mute the thread https://github.com/notifications/unsubscribe-auth/Ah3IaGTfqFEB20xn5qBmPTWPoHHgfnugks5t8oUygaJpZM4Umd_L .

-- Thanks,

Phil http://www.appaapps.com/howToWriteAnApp.html

Philip R Brenan http://www.appaapps.com/howToWriteAnApp.html

coeusprime / speech

Feature point questions, re: Janet #12