2016-08-01 ~ FGx Crossfeed Dailies Replay R1 ~ Warning: Crashes your browser #37
Comments
|
Hi Theo, Warning: just uploaded some 23 more cflogs to crash your browser ;=))
There should not be any anomoly in the lat/lon, but in reading globe-crossfeed-replay-r1.html I can see you base the tracks seemingly on the CALLSIGN... But over the course of the day, any particular CALLSIGN may have done several flights, and they could be at diverse globe locations, like KSFO, LEBL, YSSY, etc, etc... so you should not join all the dots simply according to CALLSIGN! But maybe I read it wrongly... Each CSV line commences with a FID - a unique flight ID... You should only join-the-dots according to the FID... And I would suggest you pre-process the CSV lines, and break them it up into FIDS... so you can cull the FID records even more... check if the last Now specifically looking at flights-2016-07-30.csv - it has 351,616 lines, but the first is the csv header... It contains 1,760 (unique) flights, but after the above culling, I am left with just 697 valid flights, spread over some 225 CALLSIGNS, so you can see quite a few did multiple flights... To see this in detail try - http:https://geoffair.org/tmp/flights-2016-07-30.htm#stat.alpha - it looks like most CALLSIGNS have more than one flight, during that day... A special mention about the callsign Anyway, back to your globe, I think you should consider FID as the difinitive, unique, flight marker, dot joiner... even if you do not get into culling... Regards, |
R3 is upChange Log 2016-08-01 ~ R2/R3
|
|
@fgx/owners FGx Globe Crossfeed Dailies Replay R4Change Log 2016-08-02 ~ R4Back to creating objects foe each callsigns. Means fewer objects to update 60 times per second. ;-) Will need to look at monitoring both callsigns and FIDs. Still too many zombie trails.
|
|
Hi Theo, What you choose the cf csv logs to be is your choice... I can only tell you what they are ;=)) Simply, the crossfeed.freeflightsim.org/flights.json is fetched, each 5 seconds, and stored in a by FID hash... and each flight written to a Now when the next json arrives, 5 secs later, first flights that have left are removed... from the hash... that FID will never again appear in the csv... the existing flights are updated with later info, and any new flights, FIDS, added, and appended to the So the csv list is not a list of continuous records... it is in fact a set of 5 second The blocks can be separated by comparing the
Well, that depends on what you mean... If the objects were based on FID, then when you read the next json block, you could eliminate, remove, delete ended FIDS, thus at eny one time you would have a maximum of 54 objects to update... sometimes a lots less, down to 3... This seems a better display - show only FID flights still active - gives a better, clearer, sort of by hour usage, world wide, rather than the display being massivly clutter with dead zombi flights, that will have no update... And lines between where a callsign ended the last FID flight, and started up a new FID flight, maybe half-a-world away, are very confusing, to say the least... and must be avoided IMO... And when you wrap the file, everything should be deleted, and the globe starts a new day... maybe you could add a Sun object, and move it around the globe... you already show a current UTC time... ie the CSV upd value... Your present crude next set of record reading seems to warp forward an hour about each 10 seconds... that file reading could in fact be based on a user time warp - Show 24 hours in [ 5] minutes... etc... Now I am sure you can sustain up to a maximum of 54 flight (FID) full tracks at a time... at 60 frames per second... Just an idea... Regards, |
|
Still a lot to learn here. Will get there eventually, I hope. ;-) More tomorrow or so. Got me some pretty maps to finish up... |
|
@fgx/owners FGxCrossfeed Dailies Replay R5Change Log 2016-08-03 ~ R5Still takes way to long to reload or close the page. This has to do with the trails - Three.ls lines.
|
|
Can we help each other out here.. Cut + paste CodingTheo is creating custom pages and layout and custom "theo style" cos it worked (cut+paste programming, like I used to.. To Future and making it perfectWe need to get theo to start using git.. and having a "server" running.. Updating locally and commiting will update live server also and alike... So can I suggest we make a decision on what "platform to use".. Cos we can use these dynamically as virtual machines.. We also need to automate creation of csv files and alike also.. And this is same stuff as before and the flight tracker... ie marking points in space etc.. My Opinion..Server Lang ? and I would prefer to go with revel.github.io, as a webframework, cos it simple to use imho.. biased cos a developer... but defiately easier that a lof of other "frame-works".. Next Step@theo-armour create and openshift account and a "php5 webapp".. Were not using php, but for statics... |
|
You raise a lot of really good and interesting points. Thank you for taking the time to write these down Please do remember that I am a designer and not a programmer. I happen to use a computer language as my design medium but that's like saying one person might use a sheet of paper to write a poem and another to add up a column of figures. I have jotted down some of my thoughts here: I am always open to new ways of thinking and working - so let's continue the discussion... Theo |
|
@pedromorgan, wow a whole mish-mash of things... thanks for taking the time... very educational, for me... And thank you Theo, for adding the Sandbox link, also quite edifying... Continuing the Some points observed -
Pete, like all programmers I suppose, you moved on from cut-paste As you begin to understand the process, you begin to create
A big YAY to that! As you point out it is an ongoing process... get one thing displayed, and then see if you can make it better... I am presently using my RPI2 to capture the 5 second crossfeed json feeds, and write the results to a It is not very robust, and can abort, but do not want to install say But it would be easy to add to that script an sfpt/scp copy of the last csv, to another location, on each file change... maybe as a shell/fork to avoid disturbing the current json flight 5-secs collection... Pete, any thing you can do to help this would be very, very welcome... But come on Pete Me embrace But you could host say a WHOH! As always, just IMO ;=)) Regards, |
|
@fgx/owners Just a reminder. This GitHub issues thing is like a forum. It's wise to stick to the thread. This issue is: #37 - with subject: ' blah blah FGx Crossfeed Dailies Replay blah blah' There is also issue: #34 - with subject: 'Is FGx Moribund? If so, what to do' If you ask me, we now have a STAR pattern with two issues that have converged into a path to a single runway. In other words the two issues are in the process of a mid-air collision. So, Pete, you are really good with starting fresh repos. Can you do the same with issues? |
|
@theo-armour not yet.. am busy busy.. BUT I want your space as a website to stuff.. |
|
There's likely to be two kinds of 3D: globes and maps I see you like master' and 'next'. I've been using 'build' and 'dev' or 'latest' if it's oriented to non-coding peeps. What do you like? All makes no difference to me as long as we all agree... |
|
@theo-armour, I too do not care about specific branch names, but unsure where Your creation of a HTML update, by increment the So for either the actual repo source, which as far as I have seen are usually all one branch... And for the fgx.github.io html pages, where you are using version iteration, then again, What I like about this particular globe display, is that the set of available csv files automatically grows, as each new one is pushed... using the interesting And this is from a html file in It is a little disappointing that the But after watching it for quite a while, one can get to like this like global game of golf ;=)) It takes some time to read one file. It looks like I guess at But maybe a time-warp slider - you assume daily CSV - ie 24-hours of flight records - to be player over what period of time? You could set appropriate upper and lower limits... Anyway, just an idea after watching many file replays, and not a single browser crash ;=)) Also, while you show the current Waiting for the next |
And well you should be. Currently there is no branching. I was waffling about as and when there could be branching. I find that - if you are the only one contributing - that committing from my local to GitHub host is quite sufficient. But the moment there start to be pull requests then I should and would switch to build and dev or whatever branches.
Yup. the GitHub API enables many fun things to happen
Good assumption. There is somewhere that the code is preventing proper garbage collection. I think trails is part of the issue - but this is a work in progress. One day I want to see the full day of trails. Thus creating FlyingSpaghettiMonster Type 2 (Type 1 being the STAR system )
??? Really? For me it's almost instantaneous. All the issue for me seem to relate to dumping the previous day All the other bits: we will get there... |
@fgx/owners
Warning:
Crashes your browser after your second pass through the 31Mb of data
Open in Incognito mode or in separate window
And something very peculiar is happening with the lat/lon
But the main thing:
Thank you, Geoff, for making the data available.
See also: https://github.com/fgx/crossfeed-dailies
FGx Globe Crossfeed Dailies Replay
Change Log 2016-08-01 ~ R1
The text was updated successfully, but these errors were encountered: