Notify upon specific condition

I don’t need an explicit warning as notification but I would like it, if I could set a limit for temperature and wind force, where the colour of the graph changes and an editor for these limits.
Just to be warned when I check the weather widget

1 Like

OK, I get your question.

For me this seams like one of those things that no matter how much you try to think about how you want it before you get it, it will change once you start to try to use it.

It will probably need to start with some basig guess, (like the ones you described as examples), and then evolve somehow organically.

1 Like

hi…add significant snow is on the way!

1 Like

I don’t think this is true and it counter to software development. Most software projects define specifications before starting. If you can’t describe what you want in words, I’m not sure we can randomly throw far more time trying to program something.

This is probably why I haven’t started it yet. I don’t really want to start an “organic” feature when others are not complete.

3 Likes

I might of mentioned this before but before Flowx I had several weather apps with announcements on to test them I found them annoying :joy::joy::joy: always going off large amount of time what they were supposed to be announcing didn’t happen just gave an idea of what might happen Then I found Flowx where you can scroll through the day and get all the info needed :100::100::100: no constant pinging, ringing, dinging for the same thing over and over. but then again I’m a weather geek and use Flowx no less than 12+ times per day and when possible weather events or events that are weather dependent Flowx stays open

3 Likes

Yes this is probably true. And it is so in order for developers/programmers to be able to get anything done.

I think that the “Notyfy upon specific condition”-feature is one that seam fairly simple to state, but actually is far more complex to get “right” or useful without being

, than what is easy to imagine.

There is ways to deal with this type of situations, but they are not always fast or cheap, and
it seams to me that working on other features is giving more value per spent time for Flowx at this time.

This is the true power of Flowx.

2 Likes

the reason why I said annoying is I live in a micro climate area so I get constant alerts and the storm splits and goes around. for example the last storm NWS app kept alerting for snow at my house we got a very light dusting however the north end of town got 4in. same when possible rain. at other times no alert and I get the heavy part of the storm and the Northern part gets little. not in anyway saying alerts are not useful as they very well be helpful to others. screen shots here show how some storms just go around me

https://forum.flowx.io/t/interesting-cloud-formations

2 Likes

Regarding design specification. It’s not just " in order for developers/programmers to be able to get anything done.". More importantly, it’s so you don’t months doing the wrong thing.

As such, I don’t want to start this notification feature without a good understanding of what people will use it for. If you can’t describe what you want to use it for, how am I able to design it, except for a basic system??

If I implement a basic system, the doors will flood open (I know from history) with users asking for notification A but only when C is below X. I will hack a few of these in and then find, I’ll have to rewrite the whole notification system.

For example, I designed the first graph system. Then users wanted “Q” with “G” but not “F”. I started doing a few of these and then rewrote and added the graph editor. The graphs were never designed for multiple sources or other data types, e.g., UV. Just your standard weather sources. So now I’m rewriting them to cope with other data types.

All I’m asking for are descriptions, like, I would like to know when wind is below 10knots and humidity is above 60% and there is a low chance of rain which is ideal plant spraying conditions. Check once a day (at 6pm) for the next 5 days range.

Or, if there is 20+mm of rain each day for the next 5 days which is ideal planting conditions. Check once a day (at 6pm) for the next 5 days range.

These two I made up from a email from a user I read a while ago and from my need to some planting.

It’s not too hard compared to the amount of work that goes into developing the feature. And the advantage is that you’ll have a notification system that works for you on the first go, instead of a frustrating system and a frustrated developer hearing all the conditions for notifications that should’ve been implemented.

3 Likes

I feel your frustration.

(Iterative products and systems development is what I do for a living.)

My point is that I’dont think it’s possible to “come up with a generic way to meet all needs” on the first try for a “notification editor”.

I do believe an iterative approach is needed to get this one right.

What I’m saying is that if you are not prepared for “the doors to flood open”, then it’s not the right time to start with this feature.

Notifications are a vastly different domain than displaying data in graphs and on maps, and users will react in a different way to it.

2 Likes

I agree that the first iteration won’t meet all needs - this is not my frustration. My frustration is that I asked for specific notification descriptions and I haven’t got many or enough. I made up two from a help email and my own gardening needs. So I don’t think we’ve got anywhere need the 80/20 rule, i.e., 80% of the notifications designs from 20% of the work.

This makes me think this isn’t really a big feature many want. It reminds me of when I spent 3 weeks implementing a solar eclipse path (with cloud cover) and only a few people used it. I’ve since removed the data.

Based on what we have so far, I get the feeling I could spend weeks on this, get a rubbish notification design, throw it all out and start again, except I won’t have time to start again. This kind-of doesn’t inspire me to start it. I won’t start a half-baked feature.

So I’ll wait until I’ve got a good design from a range of notification ideas from help emails and mulling it over in my head before I start it.

4 Likes

Yep, I think so too.

Sonds reasonable!

2 Likes

(First time poster)

Love the app. I was a Dark Sky user on Android and now I use FlowX daily and often. Dark Sky had a “simple” feature which was a push notification when a simple condition is met. By simple, I mean location is local, metric is singular [in my case UV], and trigger was time only iirc. So each day I would get a notification about when the UV of 4 (point at which I must wear sunscreen). I had another where temp was predicted to be 70 (at which point, I close the windows). I’d leave the notification active on my phone until one or both conditions were attended to and then swipe. During the winter, sometimes I would receive neither notification but in spring they would start up.

I’d pay Gold sub/year for that. If it already exists, I’m IN! :slight_smile:

Thanks for all your hard work.

2 Likes

Thanks for the post @DWSimmons. This is the kind of information I need.

  1. for both your cases, what time window does the condition look at?

  2. what happens when you have multiple data sources? e.g., in Europe you can have up to 12 data sources of varying resolution.

  3. what about new forecasts? e.g., HRRR releases a new forecast each hour. What happened if the temperature cycles above and below the condition? Do you remove the notification? Do you re-notify for each forecast?

I would not do push notifications since it would require a remote server, logic and maintaining a database of conditions and possibly user accounts and logins. In other words, a lot of work and maintenance.

I would save notifications on the phone and like the widget check for new forecasts and notifications every 30 minutes.

Cheers, Duane.

3 Likes

Thanks for responding.

  1. I think it was pretty dumb, i.e. check once at time of notification and spit out answer, which was totally fine for me. For the time window, that calendar day: is the event predicted to occur? If yes, when, if no ignore.

  2. I personally didn’t care. I understand from your side that doesn’t really help. Not only does a source need to be declared, as soon as such a feature is given, some user is going to want to specify which source. How about whatever source was last used to square the circle?

  3. I get where you were going on this. The notifications I desire are super dumb and more of a task reminder than a prediction and nowhere near a fact.

FlowX has all sorts of cool data and multiple sources which I barely touch. I use it for the UI/UX and the UV (bad letter mixing intended).

Another use case I can think of is kiteboarding. I don’t do it though I understand wind direction and speed are conditions that are essential to know. For example, 8 am check if sustained wind will be above 15 knots, if yes, add direction and button to open app for details, if no, ignore. I am guessing that kiteboarders have an obsession with the weather.

2 Likes

App development is funny. For some types of features, if it doesn’t exist in an app, users generally don’t care or kindly ask for the feature. As soon as the feature is added, users complain it doesn’t do A, B or F. “Give an inch, take a mile” scenario. I suspect notifications is one of those features because of it’s complexity and the range of types of notifications. This is why I worry that if I don’t get it right, then it’ll be the bane of my life.

This is why I want a canvas of types of notifications users want, then I can design dialogs to create notifications and act out what might happen. If I don’t have a broad canvas of types of notifications, it’ll become and maintenance and migration nightmare.

One approach is to limit access, e.g., to gold only, and also class the feature as super-alpha, i.e., you might lose all the notifications you setup when the app updates.

4 Likes

From what I can see we need:

  • multiple data criteria:
    – data type
    – modifier (min, max, change in 24hour, accumulation, etc…)
    – comparison operator (==, >, <, in list)
    – values
  • check period (every 30min, 1hour, 3 hours, 6…)
  • day range to check
  • time range within a day to check
  • data sources to check (all or selected set)
  • source criteria to trigger notification (one, most, or all sources meet criteria)
  • refactory period (how long after a notification is shown, shown it check again)
  • custom notification message to show, e.g., “$place_name: $name > $value $units for sources: $sources” will result in "Auckland: temperature > 20 C for GFS and Icon.

I suspect there might be a few things missing from this list.

4 Likes

I’m game to be a tester. I’m not sure what else I can help with.

2 Likes

@DWSimmons Hello and Welcome to the Forum and Thank You for Using Flowx

2 Likes

Hey,

Also super keen to see this feature implemented but understand it seems a complicated one to deliver 100% right.

I however wouldn’t need it to be overly complicated.

The following would be great:

So I guess I would select:

  • location - Bristol, UK.

  • radar - noaa gfs

  • data - temperature low limit

  • threshold - at or below 1 degree C

  • time frame - 48 hours ahead.

Notification I would see pop up:

“the temperature in Bristol in 48 hours time is set to be 1 degree C”

Hope this is the desired information to help push this a little bit closer to being added to the already incredible app.

2 Likes

@jacobgibbins Hello and Welcome to the Forum and Thank You for Using Flowx

1 Like