Improving RadarBox Automatic Database Population

[AirNav Development]

We are trying to give our best to correct all aircraft database issues even before ShipTrax is released. Unfortunately and due to the 5 projects we've talked about before we can't promise that we will be able to correct all the pending issues on time.

Anyway we kindly ask our users to, at this time, report what needs to be done regarding the database population issues so everyone feels happy and the software fully communicates with the great online database our updaters are constantly working on.

In simple terms we need concise and short descriptions of wrong behavior on the database interaction.

A similar discussion has also started inside our team.

[AirNav Support]

Just to add, here are the major issues we are aware of:

- Database must be able to handle updates of data so data is not stale.
- Allow users data not to be overwritten by data on our server
- Pictures, use most recent addition rather than most popular
- Corruption bug causing details to incorrectly have "..."

[ACW367]

Just to add, here are the major issues we are aware of:

- Database must be able to handle updates of data so data is not stale.
- Allow users data not to be overwritten by data on our server
- Pictures, use most recent addition rather than most popular
- Corruption bug causing details to incorrectly have "..."

I would change third ones to read, - photos only to be derived from NAVDATA link fields as provided by updaters or return photo not available field.  No independent search of ANet for 'most' anything by individual user.  (I will PM you seperately with how this should be working within the updater DB.)

I would change the fourth one to 'When hex not in NAVDATA, Corruption bug that does not pull ICAO data from database and instead guesses ICAO code, sometimes right, mostly wrong, sometimes blank or ...'

The other major data corruptions include:
-Corruption bug messes display of all diacritical marks   à  & ū ö etc

-Corruption bug to NAVDATA of Aircraft long name adding ... to end of name IE Gulfstream Aerospace G200...  OR Cessna 560XL Citation XLS...

- Corruption to Info displayed next to photo.  Reaches into Navdata for Regn rather than Hex

- Confusion in decode of aircraft hex leading to incrrect callsigns/heights/locations being displayed on screen
IE
   400A7C  SHT18C   LOWW-EGLL       G-DBCC  A319  bmi                  2011/02/25 12:38:01  
   400AFC  SHT18C   EGCC-EGLL       G-EUXD  A321  British Airways      2011/02/25 12:31:41

   400B04  BEE7RV                   G-JEDW  DH8D  FlyBe                2011/02/25 15:38:04  
   400B04  SHT2P                    G-JEDW  DH8D  FlyBe                2011/02/25 15:41:28  
   400B00  SHT2P    EGLL-EGCC       G-EUXH  A321  British Airways      2011/02/25 15:39:08

- Corruption bug to route if same aircraft picked up only once on seperate days with same callsign each day, displays as one flight and does not appear on reporter for latest day.
See attachments

-No 3D models or skins for hundreds of ADS-B aircraft. 3D models also set up against corrupted ICAO codes like A380 instead of A388.

- No Silhouettes for over 600 aircraft types

Every single one of the above issues has been around since V2.0 and have been reported to death in myriads of threads, so you should be well aware of every one of them.

[AirNav Development]

Our comments:

>I would change third ones to read, - photos only to be derived from NAVDATA link fields as provided by updaters or return photo not available field.  No independent search of ANet for 'most' anything by individual user.  (I will PM you seperately with how this should be working within the updater DB.)

Ok. Right now the server returns the two newest airliners.net photos for each aircraft. It saves the links on the main server database which is then passed to the client application. What's wrong with this?

-

>I would change the fourth one to 'When hex not in NAVDATA, Corruption bug that does not pull ICAO data from database and instead guesses ICAO code, sometimes right, mostly wrong, sometimes blank or ...'

Corrected. ICAO aircraft code returned directly from server and no more "guesses".

-

>-Corruption bug messes display of all diacritical marks   à  & ū ö etc

Send two examples so we can investigate further.

-

-Corruption bug to NAVDATA of Aircraft long name adding ... to end of name IE Gulfstream Aerospace G200...  OR Cessna 560XL Citation XLS...

Corrected.

-

- Corruption to Info displayed next to photo.  Reaches into Navdata for Regn rather than Hex

Explain in detail.

-

Confusion in decode of aircraft hex leading to incrrect callsigns/heights/locations being displayed on screen

Nothing can be done regarding this.

-

- Corruption bug to route if same aircraft picked up only once on seperate days with same callsign each day, displays as one flight and does not appear on reporter for latest day.
See attachments

We will check this one.

-

-No 3D models or skins for hundreds of ADS-B aircraft. 3D models also set up against corrupted ICAO codes like A380 instead of A388.

We need a full list with these errors. It's a matter of changing the sub-folder names inside the GE_Models folder.

-

- No Silhouettes for over 600 aircraft types

Where is the Silhouette team? We can add extra silhouettes for V4.04.

-

[AirNav Development]

Changes done at this time:
- Preferences form scroll box problems;
- Aircraft ICAO code returned from server and not guessed;
- If aircraft record has been manually changed by user it is not automatically updated (option for this);
- Right mouse click on grid and report problems with aircraft record for updaters team to look at;
- No more aircraft/route ending with ...;

[ACW367]

Our comments:

>I would change third ones to read, - photos only to be derived from NAVDATA link fields as provided by updaters or return photo not available field.  No independent search of ANet for 'most' anything by individual user.  (I will PM you seperately with how this should be working within the updater DB.)

Ok. Right now the server returns the two newest airliners.net photos for each aircraft. It saves the links on the main server database which is then passed to the client application. What's wrong with this?

Sorry this is nowhere near fixed.  Yes The links are populated to the updater and passed to user NAVDATA.  They are then completely ignored by the users box in the search for thumbnails for the photos folder.  This thumbnail search is still as for V2.0 and searches most popular from each individual user to populate the photo thumbnail folder.  

The NAVDATA remains the same with the central 'most recent' server link which is completely unrelated to the 'most popular' thumbnail.  Clicking on the 'most popular' thumbnail will take you to the ANET page for the correct NAVDATA derived 'most recent' photo.

See attached screenshot, this thumbnail of a previous B732 uploaded today, ignoring what is in NAVDATA links with the updater team added blocker. 

Double clicking on the bad thumbnail opens IE at the correct blocker link.
http://www.airnavsystems.com/cgi-bin/anlv_sv/photo/photonotavailable.jpg

[ACW367]

Our comments:

>I would change the fourth one to 'When hex not in NAVDATA, Corruption bug that does not pull ICAO data from database and instead guesses ICAO code, sometimes right, mostly wrong, sometimes blank or ...'

Corrected. ICAO aircraft code returned directly from server and no more "guesses".

Sorry Airnav wrong again.  Any aircraft not in the released NAVDATA will still guess/corrupt the ICAO code  when pulled from the updater server as will aircraft long names when in the NAVDATA and subsequently picked up on the box.

The last code in yesterdays NAVDATA release was OO-UAR.  The six codes below that were added or changed in the Updater DB today and pulled from the server to my NAVDATA.  For the A320 B774 and B733 the code was not direct from the updater DB but guessed correctly by my local software, exactly as has happened for these easy codes since 2.0.  The other four were not guessed correctly by my software and have corrupted. 

At the second screenshot is the five records that have corrupted in the past day with the dots after the long name.  No record had the dots immediately after I uploaded the new NAVDATA yesterday, I know, I specifically checked.  These five corrupted today as my box picked up the aircraft.

 

[ACW367]

Our comments:

>-Corruption bug messes display of all diacritical marks   à  & ū ö etc

Send two examples so we can investigate further.

In aiports table
Nagoya Chūbu Centrair Int't corrupts as
Nagoya Ch?bu Centrair Int'l

In updater DB and in NAVDATA since version 2.0, ANET photographer names with a diacritical mark have always corrupted in the PT fields
Sergio Mendes.    Becomes corrupted
Sérgio Mendes.

Also any company name with an & symbol, corrupts this to space underscore IE John Smith & Sons in NAVDATA becomes John Smith _Sons when displayed next to photo.

[ACW367]

Time for bed - Can someone post a screenshot of a TACTICAL or VARIOUS to illustrate the Airnav quote

- Corruption to Info displayed next to photo.  Reaches into Navdata for Regn rather than Hex

Explain in detail.

[ACW367]

Our comments:


-No 3D models or skins for hundreds of ADS-B aircraft. 3D models also set up against corrupted ICAO codes like A380 instead of A388.

We need a full list with these errors. It's a matter of changing the sub-folder names inside the GE_Models folder.

- No Silhouettes for over 600 aircraft types

Where is the Silhouette team? We can add extra silhouettes for V4.04.

-

The second question confuses me.  TARBAT provided in the other thread the full list of 600.  Airnav have always stated they do not require third party resource.  Therefore surely the silhouette team that created the 2.0 silhouettes was within Airnav.  If this task had been regularly reviewed within airnav over the years, it wouldn't have grown to be such a big task.

For the first question.  Can someone monitor the network for a week or so and report to Airnav every single airline code/ICAO Code combination that appears on ADS-B aircraft

Just looked at nework A332 the following airlines appear.
Those with ** are ones not currently showing in the 3D Skins folder and therefore Airnav created skins are required.  This is just a small sample.  As given above it would take me maybe a week to get you the full list of ADS-B regulars.

Network A332 at 0400Z 26 Feb
AFL
AFR
**AIC
ALK
BER
**CRK
**DAL
ETD
EVA
**GIA
**GFA
**GRL
JST
KLM
**KKK
**MEA
QTR
TAM
TCX
THY
**XLF

[tarbat]

Time for bed - Can someone post a screenshot of a TACTICAL or VARIOUS to illustrate the Airnav quote

- Corruption to Info displayed next to photo.  Reaches into Navdata for Regn rather than Hex

Explain in detail.

Hopefully it's explained with these screenshots I posted a while ago:

 

The information displayed next to the aircraft photos is wrongly derived using a REGISTRATION lookup on the aircraft table, whereas it really needs to use the ModeS hex code (the primary key) to lookup the information.

Some more explanations at
http://radarspotters.eu/forum/index.php/topic,4169.msg29969.html#msg29969
http://radarspotters.eu/forum/index.php/topic,3932.msg27958.html#msg27958

[tarbat]

-No 3D models or skins for hundreds of ADS-B aircraft. 3D models also set up against corrupted ICAO codes like A380 instead of A388.

We need a full list with these errors. It's a matter of changing the sub-folder names inside the GE_Models folder.

The first place to start is the missing models (not skins).  A quick analysis of MyLog indicates these are the most numerous ADS/B aircraft (flying over Scotland) without GE models:

B77W   65
RJ85   19
GLF5   13
ATP   8
A388   7
CL60   5
EC25   3
C25B   2
E55P   2
F900   2

Created using this SQL:

SELECT DISTINCT 
  Aircraft.AircraftTypeSmall,
  COUNT(Aircraft.ModeS) AS Counter
FROM
 Aircraft
 LEFT OUTER JOIN GE_Models ON (Aircraft.AircraftTypeSmall=GE_Models."AT")
WHERE
  (GE_Models."AT" IS NULL) AND 
  (Aircraft.ADSB = "Y")
GROUP BY
  Aircraft.AircraftTypeSmall
ORDER BY
  Counter DESC

[bearcat]

"Confusion in decode of aircraft hex leading to incrrect callsigns/heights/locations being displayed on screen

Nothing can be done regarding this."

Why not? Other boxes don't seem to have a problem. See the attached from Thursday evening, which was the correct aircraft/flight?.

[AirNav Support]

Bearcat,

We were asking about database changes at the moment. Any other issues will be worked on later on.

Also some of the changes taking place mentioned above are taking place both server side and the software side. So do not expect a difference to everything mentioned until a new software is released.

[bearcat]

This was raised in this thread, but look forward to this being sorted out in the next software release