Open the can

Well, an update for all you spam posters out there (since you’re the only ones that seem to visit this place).  Development on the app is still moving forward, but my recent career changes means I have a lot less time to devote to it than I did. 

The good news is that the app (both Simple and Pro) are as stable and functional and featured as ever.  So this is a good place to be able to leave it and make only minor improvements over time.  I’ll certainly still jump on bugs as they become known, and I’ve got a few more good ideas up my sleeve on features to add…but it is probably 85% complete as far as what the app will be for the rest of its life.

Keep the comments (from within the app/troubleshooting) coming, there are lots of good ideas and I love hearin’ em.

Well, I can’t help but be a LITTLE pleased.  There’ve been over 20k downloads (though only 14k still installed) of the Simple Weather Alert app, and now 500+ active users of the Pro Weather Alert.  That’s pretty cool.

The app has slowly gotten better.  Today was pretty big because I finally nailed down some code that had been a real sore spot.  For some reason my timers kept getting killed.  Well, it wasn’t the timers really, it was the code that was supposed to set the timers.  I’m still not entirely sure why, but apparently just trying to catch the Exception wasn’t good enough…I had to explicitely catch the IOException.   Now THAT’S annoying.

Anyway, I feel 1.4 is just about as stable as it can be.  So now it REALLY REALLY is time to jump into the GPS alerts.  I think I can finally do that now without having anymore raging fires to put out.

Ok, the “night of endless alerts” appears to have passed.  The NWS has now finally gotten their alerts to come up on web browsers again, and I assume that also means they’ve stopped playing with the ID value of the alerts, which caused a new alert each time.

With the addition of version 1.2.1, which added a lot of stuff (and 1.2.2 for Simple Weather Alert, which added stuff that was supposed to be in 1.2.1), it’s finally time to move on to GPS/roving alerts.  I think this will be fun.  I’ve created a new app which is basically Pro Weather Alert, but with the GPS/roving code in it.  I plan, once it’s functional, to offer that free on the market to get feedback on how well it works.  That particular app will only work through June and then it will die.  Once I’m pleased with the location functionality, I’ll incorporate that into Pro Weather Alert.

If you’ve been experiencing the “constant alerting” issue today, know that the problem lies with the NWS xml feed.  It has been a problem all day, and I’ve emailed them about it, but it persists.  I’m going to try to figure a workaround until they fix the issue so people don’t continue to get constant alerts.

We apologize for any problems.

Wow, I don’t need to look at the national forecast to know that there’s hardly any bad weather going on today…. I know just because of the lack of emails, comments, or debug data on the app.  Downloads of the free app have dropped off today too.

And it feels like all the air has gone out of my tires.  I just sat kind of blankly staring at the screen for a while last night trying to figure out what to do next.  I’ve been very much in emergency-fixit mode the last few days, so to suddenly have a lull was a bit surprising.

I *think* the app is better now. I made some changes to how the timer is reset, and to how the alerts list is queried.  It’s a little more brute force, not as elegant, but at this point I’m going for “it works”. 

So what are the major lessons learned here? 

• I should have put out the free version and worked out the bugs before putting the pro version out, or have put out a *beta* version first.  The fact that it didn’t work when people needed it to caused my rating to get hammered.  I tested this app for weeks…but my test environment, emulator or not, is simply no match for being tested simultaneously by 5k people in a real-world environment (duh).  I should have fully instrumented the app and given people a “send data” button when things went wrong.
• People don’t like the in-your-face “buy the pro instead” part of the simple app.   When there were limitations (couldn’t add a new location, couldn’t see all alerts) it would say to buy the Pro app instead.  It is perfectly fine to have limitations in a free app, but they probably shouldn’t be so blatant.
• People don’t like multiple updates a day.  This was definitely my fault, and related to the buginess above.  But one particular update, I KNEW I had a bug in the code, but before I got back to my PC I had forgotten about it and published the update.  DUMB!  I had to do an “emergency” update and beg people to fix it so they could get notifications again.  My previous experience with writing code for individuals usually involved like 5-10 people, so zipping out a new, fixed version was no big deal.  When you’re talking this many, you really need to control what and how you put stuff out.
• People REALLY don’t expect to hear back from the app developers.  The first emails I got, when I replied to them, they were honestly shocked that I did, and how fast I did.  Well, I’m “lucky” to have the time right now to sit here and wait for emails (while adding to the app) so that’s why I’m so responsive.  But I LOVE getting emails about the app and responding with people.  In the future (I hope) I won’t be able to respond so quickly, but I certainly still will respond.  In the future future, I hope sitting here respond to emails about my apps is ALL I do!

I don’t regret the experience at all so far, I’d just say I was pretty naive.  But then, that’s kind of expected when embarking on a completely new experience.

I completely understand now the lure of games on Android phones.  No, not PLAYING them…developing them.  Why?  Because no one really cares if your game sucks, or breaks, or simply doesn’t work.  Sure, they may get angry and give you one star…but who the heck cares, it’s a GAME!?? 

Now, attempt to create a weather alert app, and you’ve signed yourself up for a lot of hurt.  People expect it to work, DUH.  But even more so than, say, a tip calculator app…you feel a bit of pressure for the app to do what it’s supposed to.  Not to be too dramatic, but people’s lives really are involved.  I’m certain anyone living in the midwest has learned to deal with tornados and warnings long before my app came along, so it’s not like I’m suddenly crucial.  But still…the one time someone gets a notification from my app when they couldn’t have gotten it from somewhere else…that’s awesome.  And the many times I’m hearing about where it DOESN’T do what it’s supposed to?  That’s sucksville.

I pondered many times today just yanking the app and refunding everyone.  That’d be alot of clicking, but still…then I’d be free….FREE!!  But I’m not going to give up yet.  As long as I still have the time to devote to this, I’ll make every effort to get it to work flawlessly.

And in the meantime, I daydream about a game….a fun game, where people *laugh* when they play it….where no one can die if it Force Closes on them, and at worst they give you 1 star and uninstall it, but hey, it was free, so go for it…. *drifts off*

So, the honeymoon is over and the stars begin to fall.  

It is oft criticized of the android market that the comments are a one-way street.  The customers can post anything they want, nice, evil, or simply wrong….and the developer can’t say a word in reply.  At first I thought this was rather unfair, and counter productive.  When a customer actually asks a question in the comments…how am I supposed to answer??  I can’t change the product description unless I post an updated version of the app, so no luck there.

I now think that the one-way street is actually a good thing.  I’ve been in enough forum flame wars to know how quickly the comments section would devolve.  Even simply replying to someone who complains that they have to pay to see the most important alerts (which is not true)…I don’t think I could do it without being just a tad nasty.  And then it would just get worse from there.

In many ways I just have to stand and take it.  I honestly don’t blame people for rating it low when it doesn’t work.  That’s perfectly justified, and the remedy is completely on me.  But I do hate them rating it low because it doesn’t do something they want it to do, when that was clearly spelled out in the description.

I’m not going to chase the stars in the market.  That’s what Super Mario Galaxy time with Sami is for.

We’re into day 2 of Simple Weather Alert and Pro Weather Alert being released.  So far I’m seeing 1800 downloads of  SWA and 42 of PWA.  I have NO idea if that’s good or not, but I’m pretty pleased.

Reviews have been slow to come…I think few people take the time to do that, and those that do usually do so because they’re angry.  I’ve been lucky though, since I’ve had 6 5-star reviews of SWA so far.  I finally got a 1-star rating today.  “Deb” was disappointed that she could only use one location.

I did ponder that one for quite a bit when designing.  SWA is just as USEFUL as PWA, it’s just not as generous.  You only get 1 location to monitor, and in that location, you only get to review the top 2 alerts.  Any given location doesn’t often have more than 2 alerts, and if it did, those alerts are usually of Minor severity or less.  So you’ll still always see the important ones.  But that’s also part of why I restricted it to 1 location.  If I allowed 2 locations, then it becomes a juggling act of what alerts to make viewable and which to not.

Of course, those limitations were completely designed to urge people to shell out the $2 for Pro Weather Alert.  I had originally toyed with the idea of limiting even PWA to only  5 locations, but in the end, I think you really have to make the “Pro” app above and beyond everything else…unrestricted basically.

Even at 1 star, I appreciate the feedback.  I’ve received several emails asking about specifics and how to configure the app, and I enjoy those very much.  It’s all to the good.

Only an hour live and already have an error report!  And I’ve already posted an update too…so that’s not so bad.  At least, I HOPE I fixed it.  Man, it feels hard to go to sleep, because I’d hate for the app to suddenly fall apart and everyone trash it in the reviews before I had a chance to fix it, but *shrug*  what can ya do.

I’m building my list of planned upgrades/improvements to the app already.  I think that’s important to do…it’ll make for loyal customers (who hopefully tell friends about it) and make the app more appealing to new customers.  And some things I definitely WANTED to do, but it was time to get the app out there.

So far I have:

• Allow for sending of alerts to friends (I might add that to both pro and simple, not sure yet)
• Put in a preference for if every update to an alert is considered a new alert, or only new events.  This one is hard because the NWS does not put a specific alert ID on each alert….so there’s really no way to know if a new alert is simply an update to an old alert, or really a new alert.  But I can at least check and see if that particular event (Tornado Warning, Flood Watch, etc) is already present, and then not notify on that basis.  Either way, if I leave it up to the user, then they can deal with it.  That’s for both Pro and Simple
• Keep expired alerts, with automatic or manual purging of them later.  Pro only.
• Put in GPS-guided alert zones.  Pro only.
• Have a little more “grace” when people put in places and locations…so, if St Paul fails (which it will), then I should automatically search for Saint Paul for them.  Or, if a direct search fails, I should probably convert it to a LIKE query and try again, and then the user would see “Did you mean?” and give them what comes back.  That’s a bit involved, but over the course of 200+ queries in the database, I’m already seeing 12 failures.  Most revolve around using ST instead of SAINT (which the error message DOES say Don’t use abbeviations), but some are mis-spellings too.  Then again, I don’t know that a LIKE query will fix that.
• I do need to return to them that a location does not exist, as a specific error message, instead of the generic “There was an error”.
• Let users select a ringtone for the emergency and urgent levels, instead of just using default. (Pro version)
• Change the alert icon in the status bar to red, and yellow for urgent.   Can’t do this, it goes against the status bar icon guidelines set forth by android (http://developer.android.com/guide/practices/ui_guidelines/icon_design_status_bar.html)

We’ll see what else gets in my mind to do.

With one final effort, we’ve now published our FIRST android apps….Simple Weather Alert and Pro Weather Alert.  The guts behind these apps is from some previous work I did with the NWS alerts feeds, and I thought  it’d be useful to bring those alerts to people’s phones and tablets.

My sincerest hope is that the apps actually WORK and they some day provide a very pertinent and useful alert to someone.