[James Ross]

Eldorado.Railroad, on 27 December 2023 - 09:40 AM, said:

Somewhere, whether on this site, or elsewhere I have seen a graph of O/S used with Open Rails. Why would you want/need even more telemetry WRT O/S running Open Rails?

We do not know what operating system people use with Open Rails. We only know the version number of Open Rails itself from the update checks, and we'll not be adding anything else to them.

You might have seen OS or geo data for downloads from Google Analytics on the website. I can't think where else it would have come from.

Eldorado.Railroad, on 27 December 2023 - 09:40 AM, said:

To my knowledge certain users of Open Rails (or ANY other O/S and/or programs) eschew even more tracking/spying/"telemetry"/"surveillance capitalism" and last but not least, SaaS. To what domain EXACTLY is this user information going to?

All the code will be public, including the server-side collection service, but none of that has been written yet. The domain will be "telemetry.openrails.org" to keep it separate from everything else.

Eldorado.Railroad, on 27 December 2023 - 09:40 AM, said:

I have a severe skepticism about anonymous data collection. Short of using some kind of VPN/TOR network IPs can and will be collected. With that, comes geolocation of where the user resides. I don't think I would feel comfortable broadcasting what equipment I am using in Open Rails to the world at large. Heaven forbid that a precise geolocation would give the unscrupulous the knowledge of what hardware there is to steal.

Unless you have disabled the update checks, you're already providing your IP address, as it is inescapable when communicating over the internet. We only use it to count unique users, however, and have not geo-located anyone with it.

We will be taking extra steps (now documented in the OP) to prevent linking your IP address with your telemetry data, should you choose to opt-in.

Eldorado.Railroad, on 27 December 2023 - 09:40 AM, said:

Even if the desired outcome as outlined above is achieved are we once again in a hurry to cast off as many users as possible by killing Win 7, however "small" that number may be?

Windows 7 is the limiting factor for a few things people would like to do, like upgrade the version of .NET, but ORMT is not aware of a compelling reason to drop it for now.

Instead, we're more interested in the Direct3D Feature Level minimum being increased from 9_1 to 10_0.

Eldorado.Railroad, on 27 December 2023 - 09:40 AM, said:

I doubt very much that you will allow anybody to audit exactly what raw data is being collected on the chosen destination domain. Personally, I think this raw data collection should be transparent for anyone to see, otherwise don't do it. With this kind of telemetry you invite users to block access to the collection site via the router as a precaution. Given that Open Rails already notifies a user that an upgrade is available, blocking access to the domain that does the notification might not be the desired outcome.

All code for collection will be public (inside Open Rails itself as usual), as will the telemetry collection and processing webapp. As will be clear from the code (once it is written), the data is identical to the top of a log file (minus the date/time).

Likewise, the processing and analysis of the data will be public (inside the webapp).

However, the raw data is not going to be public, since there is still a chance that it may uniquely identify users. We won't know if this is the case until we actually collect the data, though, but we don't want to accidentally reveal something by starting off with all data being public.

Eldorado.Railroad, on 27 December 2023 - 09:40 AM, said:

What price progress?

Unfortunately, it appears impossible to provide glTF support (which would provide a huge upgrade in graphics for new content) without also dropping the DirectX 9-level hardware. The graphical updates are desperately needed as shown elsewhere in this forum.

There is no need to drop Windows 7 for glTF, though, and since Direct3D Feature Level 10_0 graphics card started 2006-2008 (a few years before Windows 7 in 2009) I am hopeful that everyone on Windows 7 will already have a 10_0-capable graphics card.

[Eldorado.Railroad]

More comments



Above is the "discovery" of how the "Update" feature was used in Menu.exe

No where that I could find was there an indication in the OR manual or on the "Update" page in Menu.exe that the request for possible available updates was being used to track users preferences in terms of what version of OR they were using. In the year 2023 this type of alternate meta data gathering is at best suspect.

Additionally on the https://www.openrail...ver/statistics/ page the methodology was given as to how the data was "biased":

About these statistics

"Monthly Users" and "Monthly Versions" count up the number of unique IP addresses* seen by the update server* over each 28 day period*.

This will under-count shared internet connections all using Open Rails, and over-count users on frequently-changing IP addresses.

Open Rails checks for updates once per day, but only when opened.

Rather than actual months, which vary in length and don't align to weekdays/weekends.

It is also clear that the methodology used to create the data used in the graphs has at least one huge flaw, possibly misleading IP addresses. There is nothing stopping a user or group of users using a TOR network proxy to produce IP addresses that have no correlation with a unique user/location. Yes, but is was stated that the purpose of collecting more meta data was not to aggregate IP addresses, but it happened anyways.

In this case, to display the data in a sexy fashion, a third party is used, that uses the data off the OR network (.csv), through plotly.com. How and why should they be trusted? Heaven forbid that the wrong .csv file be connected to plotly when displaying that "manipulated" data.

There are two factors that must be considered BEFORE the contemplation of extracting yet even MORE of the users data.

These are:
Moral hazard (https://en.wikipedia...ki/Moral_hazard)
and
Information asymmetry (https://en.wikipedia...ation_asymmetry)

I suspect that a simple mea culpa will not do. It is fantasy to think that any connection to the internet is not under constant attack. You are going to be very lucky to have an exposed server that collects any data not be subject to constant invasive attacks. This type of collection could be subject to "man in the middle" or worse attacks that will either compromise you or your users. Why would you want to do that?

It is clear from the https://www.openrail...ver/statistics/ that there has been an information asymmetry. Which means the ORMT has been collecting IP addresses and with it the possibility that either the ORMT or any person who was "trusted" with said raw data with ORMTs blessing have been using that data for purposes other than described in the OR manual. In addition while you have collected those IP addresses the users have NOT been privy to the personal, read: at home, IP addresses of the ORMT themselves. From a privacy perspective, how is this a good thing? This is the kind of asymetrical "privacy" that google/microsoft/et all engage in. It is left as an exercise to prove beyond any doubt that there is no moral hazard here. With the public exposure of how data was collected, without the users consent for the purpose demonstrated, on the https://www.openrail...ver/statistics/ web page you have already exposed that there is a problem, even before the ORMT attempts to aggregate even more meta data from its users.

It is rather deceptive to preface all of this as being "opt-in". The user will have no clue whatsoever how the data will be used, in any future context. Nor is there any guarantee that there will not be some ulterior motive to this "opt-in" meta data aggregation. Some future ORMT principle may have a different perspective as to how to use this meta data. That perspective may not be as benign as you have alluded to. Once the meta data is collected there is no impediment for misuse.

Really the need to have more data is not really a problem with updating/upgrading OR. You do not need the extra meta data. Instead I suggest that ORMT be less timid and move forward with certain things, such as:

Update the DirectX requirement from 7/9 to 11/12. Windows 7 supports Directx 11/12. You would force your users to update their GPUs towards DirectX 11/12 compatibility.

ORMT will have to demonstrate with a "proof of concept", in concrete terms, such as a working runactivity.exe that proves that DirectX 11/12 makes a huge difference in graphics rendering/performance with Monogame. This implies that it puts the work of any development thus far to shame. That the results are not incremental BUT better by orders of magnitude. The lack of proof just imparts a perception of "gaslighting". Change for the sake of change only. Do you own stock/equities in Microsoft? If you are developing software for a living and are forced to update to the latest Visual Studio, I understand your motives. That does not mean that OR has to move along with little benefit to show for it. You might just need more than one hardware platform. You are not alone!

People who have slaved behind a terminal/monitor/laptop etc know with certainty the magnitude of any change to this project. ORMT should re-iterate in plain terms just how difficult making a route editor is going to be. Ditto with the fantasy of jumping ship to anything like UE5.

As I have indicated before, the "smarting" of OR forward due to the "hickory stick" imposed by Microsoft's insatiable need to create a steady revenue stream is not a method to conduct this project/hobby. There is a multitude of "sub standard", by todays "standards", MSTS content. That content has yet to be replicated yet alone updated in its entirety. It is in use every time runactivity.exe is executed. Perfectly usable OLD data.

Happy New Year 2024

Steve

[Weter]

Thanks, Steve; so to You.
Don't You think, they have been manipulated from outside then?
"...for they know not what they do..."

[Csantucci]

I won't go into details. I trust James, both technically and from a human point of view, and therefore I'm sure he will do a good job on this, without damaging people's privacy.

[cjakeman]

Csantucci, on 01 January 2024 - 01:45 AM, said:

I won't go into details. I trust James, both technically and from a human point of view, and therefore I'm sure he will do a good job on this, without damaging people's privacy.

Thanks, Carlo, for that vote of confidence in James.

Eldorado.Railroad, on 31 December 2023 - 02:41 PM, said:

No where that I could find was there an indication in the OR manual or on the "Update" page in Menu.exe that the request for possible available updates was being used to track users preferences in terms of what version of OR they were using.

Good point, Steve. I will update the Manual.

[Bert Wise]

Just curious where OR New Year MG shows in the chart of versions?

Thanks

Bert

[James Ross]

Bert Wise, on 01 January 2024 - 03:07 PM, said:

Just curious where OR New Year MG shows in the chart of versions?

OR New Year MG is not an official version so it does not get updated through the openrails.org website, which means we can't track versions via update checks.

The opt-in telemetry is separate from the update checks, however, so if Carlo chooses to keep it available and sending data to the same place as official versions (i.e. telemetry.openrails.org), such versions can be counted there, for people who opt-in.

[Csantucci]

Hi James, thank you very much for the offer. I'll check what complexity that requires to me to develop it.

[Genma Saotome]

Eldorado.Railroad, on 31 December 2023 - 02:41 PM, said:

[
I suspect that a simple mea culpa will not do.
Steve

It is pretty clear it won't do for you.

Steve, if OR was a web app instead of on your PC, all the above would have been captured -- and more -- because that's how things are. I hate it, you hate it, everybody hates it... because it is always been used to sell all of us to someone so they can try to get our money. Except for OR. They've got nothing to sell and as long as they pledge not to share any info outside of the project there is no offense to worry about. Instead they're trying to make the software better. IMO that is very reasonable.

[FS.E652 091]

I'll get straight to the point without mincing words..
How many apps we install on the PC ask to share telemetry data to improve themselves?
from the moment we download an app or a game, by clicking on share game information, we do nothing but share our telemetry data so that developers can have more data available to better develop that application\game..

Therefore, from the moment in which I totally support the development of OR and I believe in the work of the development team (from the moment in which I have always been at the forefront in creating scenarios to be distributed to schools, associations and also to people in the railway ) I totally support and trust the team's choices, as long as the simulator is developed at 360 degrees both graphically and technically, we are in 2024, I myself am upgrading my PC by installing performance components and a 12 GB graphics card, for me performance is important, just as a well-done program must be.
The old MSTS and the old low-performance PCs are now (at least for me) stuff left in the past and which have taught us to progress..! but now we need to change course and so
Hats off to the team, and you have my support!