Sentinel Database Update & Favorites?

[ofd8001]

Ok, I'll bite for those of you who insist that this can be done. The Minnesota statewide trunked system is called "Allied Radio Matrix for Emergency Response (ARMER)". I shortened this system name to "MN ARMER" because the original is too wide to fit on my 436 screen.

How would Sentinel know to check the Minnesota system for updates because I've changed the name? (Me changing the way I do my Favorites Lists, just so's some folks can get their stuff updated doesn't seem too fair to me.) Also, wouldn't the program go nuts trying to find a system that I've renamed? Would it generate an error message, then lock-up or just crash?

Next, I organize my Departments based on service - as in all fire channels in one department, all police in another. How would the revised program know which to check, especially since I've changed a lot of the channel names. Will I get duplicates?

As a guy who did some software work back in the days of Microsoft FoxPro, I see this as being a lot more difficult and time consuming for the ultimate gain.

With all that said, I'd welcome someone to actually do this and prove me wrong.

 

[Ubbe]

It looks to be impossible to update FL's that you have edit to look differently in any respect.

The database has a unique reference number for each frequency, site and TG and when the database gets updates it takes that updates reference number and write the new data to your own local databases reference number. Uniden choosed to not use reference numbers in the favorite list so it is impossible to update that info as you could have changed a name tag or other info that would make it difficult to match where any new data should go.

A program for updating FL would require that no data have been altered by the user, but then you could just as well update the DB and then simply export to FL once again.

It would be easy for Uniden to change Sentinel to keep the reference number from the DB in the export to a FL and then have the update function of the DB to check all FL's for any reference number that needs to be updated, and display on screen what the old and the new data looks like and ask if you want to overwrite your old data. But for some reason that feature was not considered important to the user.

/Ubbe

 

[ProScan]

If going the Sentinel route, as you know, we don't know if Uniden plans on any more Sentinel versions But let's say they are.

What they could do is use the indexes in the HPDB (AgencyId, CGroupId, CFreqId, etc.) . I'm assuming that they are the same indexes that RRDB uses (cid, scid, fid, etc.) (I haven't tested it but it really doesn't matter) and copy those indexes to the Favorites reserved fields in the tags (Conventional, Trunk, C-Group, T-Group, Site, C-Freq, T-Freq, TGID, UnitIds)

That way Sentinel can compare the Favorites against the HPDB using the indexes and display the differences (missing, changed, & additions)

The only problem is when writing to the scanner, if a system is edited in the scanner, just one channel, the whole system is reformatted and the reserve fields may contain blanks after reading. Problem can be overcome by a helper function and file to reconstruct the indexes after reading.
{edit} I like to add mainly if Uniden is reading this is a firmware upgrade that keeps the reserved fields intact when doing an internal reformat or make then "user defined" so what goes in, comes out.

 

[SChuck]

You can update Favorites list from Sentinel. Right click on any field in the database and choose "append to favorites list" choose the list and it is done. Does not work for sites, you have to update those manually. If you have made custom names for your lists it will add a new tree to the list.
I find it best to update my lists manually if they are small. I think if you have the time to read and post to this forum you can find the time to keep your list up to date.
I know that it can be a pain with some systems with long titles like here in Michigan but if you want the ease of easy up dating leave them as they are in the database. As with all things computer software wise you need to experiment with what will work or not work.
When I travel here in Michigan I have a favorite list with the full MPSCS system and use location control, it works great!

 

[ofd8001]

It would be easy for Uniden to change Sentinel to keep the reference number from the DB in the export to a FL and then have the update function of the DB to check all FL's for any reference number that needs to be updated, and display on screen what the old and the new data looks like and ask if you want to overwrite your old data. But for some reason that feature was not considered important to the user.

And then how about channels I have deliberately omitted from my Favorites List? If this "new version" was to examine my Favorites Lists and suggest adding channels that are not there, it might force me to add something I do not want.

I'm as lazy as the next guy and like to see stuff automated to the extent reasonable, but I don't want some software trying to make me do something I'd rather not. I might end up spending more time undoing something than it would take me to add it manually if there was a change I'm interested in. It could also add a lot of time asking me if I want to add a new channel that isn't there.

 

[jonwienke]

No need to query the user. If the global is set to update existing entries, say 100, and 15 are individually locked, then only 85 would be auto-updated.

The user would still have to individually specify which 15 changes were "locked", and which 85 to auto-update. The software can't automatically make that determination.

 

[ProScan]

And then how about channels I have deliberately omitted from my Favorites List? If this "new version" was to examine my Favorites Lists and suggest adding channels that are not there, it might force me to add something I do not want.

I'm as lazy as the next guy and like to see stuff automated to the extent reasonable, but I don't want some software trying to make me do something I'd rather not. I might end up spending more time undoing something than it would take me to add it manually if there was a change I'm interested in. It could also add a lot of time asking me if I want to add a new channel that isn't there.

Take a look at visual file comparison software such as ExamDiff and WinMerge, and see how they show the differences.
The UI could be similar in showing the differences. The left pane shows the system from the HPDB and the right pane shows the system from the favorite. If one side is missing channels then that side will show placeholders. If one side has additional channels then that side will show a certain color. Changes will show in another color. Of course instead of raw text, the display should look more user friendly such as the Service Type showing instead of just a number.

 

[ProScan]

I envision the flow to be something like this.
1. Click the analyze button.
2. Show all systems in all Favorites.
3. Color code and label the systems that are identical and mismatched.
4. Click on the mismatch system. The left pane shows the system from the HPDB, and the right pane shows the system from the favorite.
5. Some type of an editor on the right side (favorites) that can delete rows, add rows, and make changes.
6. Click the Save button then go to the next system that's different from the HPDB.

 

[ProScan]

This can and should be an open source project. Have to have a proof of concept.

 

[jonwienke]

I envision the flow to be something like this.
1. Click the analyze button.
2. Show all systems in all Favorites.
3. Color code and label the systems that are identical and mismatched.
4. Click on the mismatch system. The left pane shows the system from the HPDB, and the right pane shows the system from the favorite.
5. Some type of an editor on the right side (favorites) that can delete rows, add rows, and make changes.
6. Click the Save button then go to the next system that's different from the HPDB.

THIS. Which is obviously not a "single-checkbox" user interface. Each difference has to be audited by the user to decide whether to ignore, update, add, or delete.

 

[belvdr]

The user would still have to individually specify which 15 changes were "locked", and which 85 to auto-update. The software can't automatically make that determination.

This isn’t difficult. The default would be to adhere to the global setting, just like the filters. It wouldn’t be a prompt for every change on every weekly update.

 

[ProScan]

THIS. Which is obviously not a "single-checkbox" user interface. Each difference has to be audited by the user to decide whether to ignore, update, add, or delete.

Perhaps options something like this
[X] Keep User Text Tags
[X] Don't Add Missing Sites
[X] Don't Add Missing Groups
[_] ...

Do This For All Conflicts

 

[jdolina]

I had started down this path Free Software: BCDx36HP Favorite Helper
But did not have time to keep up on program as it had limited use.

 

[jonwienke]

This isn’t difficult. The default would be to adhere to the global setting, just like the filters. It wouldn’t be a prompt for every change on every weekly update.

You'd still have to tag every mismatch with whether to follow the global setting or an exception, which would pretty much defeat the purpose of having a global setting.

I spent over a decade in the IT business, and replicating data between databases was a big part of that. It's never as simple as you're trying to make it out to be. It's clear from your comments you've never done anything of the sort.

 

[belvdr]

You'd still have to tag every mismatch with whether to follow the global setting or an exception, which would pretty much defeat the purpose of having a global setting.

I spent over a decade in the IT business, and replicating data between databases was a big part of that. It's never as simple as you're trying to make it out to be. It's clear from your comments you've never done anything of the sort.

So what? I’ve been in IT data for 26 years. I do this all the time, with various database systems and so-called NoSQL systems. It’s called a default. It’s very common in databases.

Feel free to disagree, but I don’t feel you have an understanding of what I’ve stated multiple times.

 

[jdolina]

My original thought was to have a setting for each type of record store . i.e. Trunk channel (channel name, avoid tgid service type delay....etc.) for the fields in the record you would check whether you wanted that field say 'channel name' updated if different. Then you would need to have a setting to delete records in FL not in update or vice versa. The settings might be global or by FL.

 

[Ubbe]

I don't know why Uniden bothered to add firmware code to strip the favorite lists of all index numbers. The same way that all parameters in the hpd file for a favorite list, alerts&lights and delays and volume offset and whatnot, can be copied and inserted into the blank spaces seperated by tab in the database hpd file and will use all those settings while scanning the database, the same also works in reverse.

I copied a chunk of the database with all index reference numbers and used in a favorite file, and the favorite file already have blanks at the positions where the index numbers should be, and it works and scans as it should. But if I edit something in the scanner the firmware deletes all index numbers from the whole favorite list. That's totally unneccesary and only adds code to the firmware.

It must be the same firmware code that are used to get the information from both the database and favorite lists, they have to use the same format, and scans based on that info. The format for hpd files are fixed and cannot be broken, the positions for the index numbers has to be there but they never interfere with scanning as they are only used when doing updates. Uniden could easily remove some bloat code and clean up the firmware and add selections to do favorite lists updates, at the same time as they restore the "compare to database (and favorite lists)" function that stopped working years ago.

/Ubbe

 

[jonwienke]

So what? I’ve been in IT data for 26 years. I do this all the time, with various database systems and so-called NoSQL systems.

Then you of all people ought to understand that in this context, a default isn't very useful. If you set the default to preserve the status quo on the favorite list side of things, you're essentially disabling replication. If you set the default to replicate changes from the main database to the favorite list, then manual edits to the favorite list will get wiped out during replication. Regardless of which option you choose, it's going to be wrong too often to be useful.

 

[belvdr]

Then you of all people ought to understand that in this context, a default isn't very useful. If you set the default to preserve the status quo on the favorite list side of things, you're essentially disabling replication. If you set the default to replicate changes from the main database to the favorite list, then manual edits to the favorite list will get wiped out during replication. Regardless of which option you choose, it's going to be wrong too often to be useful.

We will just disagree on this one.