Update Notification in Java apps

[Andy Wilbourn]

I have the desktop app I created and would like to have it notify the user if there is a newer version when it launches, as most good apps do these days. Since our site is mostly a WIKI, how would you propose this be accomplished? The only thing I can think of off the top of my head is to have a single page labeled as current version for the tool and the app will make a call to the page and parse the HTML string to see what version is there and if it is greater than the version running. Normally I would create a web service of sorts to do this, but trying to figure out how we can do this with the ability to have a web server app running as well.

 

Just interested in ideas if there is something better than what I have come up with. I am not familiar with the Google hosting to know if there is some sort of coding they do support for apps and we could have a config file or database storage of the versions so when parameters are passed to the page it will respond with the appropriate data.

[Steve]

That's probably the way I'd go about doing it given the circumstances.  One thing I'll mention...if it were me, along with notifying the user of the update, I would let them know what version of MLS the update supports.  That way they can decide if they should accept the update based on the MLS version they are running.  Ideally they would update MLS as soon as possible, but unfortunately that doesn't seem to be the case.

Steve

[Andy Wilbourn]

Well I had the same type of thought to version the metadata of  MLS, but Ross some time ago didn’t think there would be a need for that and since he was managing the database I dropped it. So I was left feeling he will handle all the changes needed in the DB. I still contend I need know the headers and map them appropriately, as the app could handle massaging the data to allow it to keep working, perhaps just won’t have whatever data is new.

 

I have built the converter with Stake files in mind and multiple languages. I have some clean up for some of the popup messages to not be hard coded, but to come from the resource file to be more ready, but that shouldn’t be a lot of rework. The whole app probably needs to be reworked to follow java conventions, as the package name does not follow those I have seen (what can you expect for my first Java app J).

 

I can’t believe I have had such a hard time trying to find out how to set the image in the top left corner and thinking that is the same for the taskbar in Windows. The default Java icon is not the best for an app. You familiar with SWING at all as to how to accomplish it? I have grown a real appreciation for the Visual Studio tools I use on a daily basis. I guess I am just not looking in the right place to find the answers I need. I find blurbs here and there, but then it is something else not what netbeans used for their base app which is a FrameView.

[boomerbubba]

Andy,

It is true that we never have evolved a methodology to deal with
backward- and forward-compatibility issues of MLS export formats.
That really is quite a complex problem. I have not solved it, but
only suggested that it might be solved partly in the SQLite layer.
For the foreseeable future, we only support one input file format.

But it seems to me that the baseline functionality that is still
missing in the converter is simply validating the basic format of the
input files, or at least their headers, and notifying the user if
those CSV input files are invalid. So, for example, the process
should abort with a graceful error message if the user attempts to
submit stake MLS files, or MLS files that fail to include the
"Preferred Name" field.

--Ross

[Andy Wilbourn]

I will publish a version to prevent Stake issues, but feel we really should have a Stake version. I know you have not been able to get your hands on the files, so that may go by the way-side as the new mls online that you learned at the conference may fill the gap.

 

I disagree with the app not doing anything with the files, but I do agree that it should prevent someone from thinking they can convert if something is missing. This is where knowing the possible columns based on version is important and something I wanted to store in the database.

 

Anyhow, I will look at the new beta stuff and may make sense to jump on board there to have influence and make apps work in disconnected mode.

[boomerbubba]

Andy,

The sad fact is that I have been stymied in getting access to stake
MLS test data, so I am no longer optimistic that will happen anytime
soon. The issue got forwarded to Salt Lake, and while no official
would expressily say no, no official would say yes either.

And that matter got conflated with the question of whether the Church
will sponsor its own Ward-Tools-like project. (There has been no
official announcement to that effect, but my understanding is that
such a decision could come soon.) If that should happen, the larger
question is whether there is a rationale for third-party apps such as
Ward Tools, or whether open-source developers should redeploy
themselves as "community" developers under the Church umbrella. I
have already signaled my own willingness to participate in such an
official project.

There are even speculative rumblings that the MLS exports would
disappear altogether. I hope that never happens, because the generic
export functions are very useful to local leaders and clerks for many
reasons having nothing to do with dedicated handheld apps. I started
a recent thread on that topic in the LDS Tech Forum:
http://tech.lds.org/forum/showthread.php?t=4933

In any case, in the short term, I can't expect to have a stake version
of LDSql. I would really like to for generalized reasons so long as
MLS exports exist. But the most urgent demand for this enhancement
has been to support Ward Tools development, where I know it is
critical path. But if the Church does declare that it wants to occupy
that space, the data feed and test data would come from expanded
online APIs that would likely be available only for official Church
projects. Meanwhile, LDSql remains generally useful so long as MLS
exports do exist. If I had no time constraints of my own, I coiuld
fall back and generate emulated stake MLS test data completely from
scratch, without it being seeded by actual MLS exports at all. But
that is a non-trivial task and I don't have unlimited time..

--Ross

> > > it will respond with the appropriate data.- Hide quoted text -
>
> - Show quoted text -