Talk:FreeTrack

TIRServer.dcu
The citation to the FreeTrack source repository lists TIRServer.dcu in binary form and provides no source file. However, we need a citation to documentation or other information indicating that TIRServer.dcu is FT's code, is used by FT, and therefore FT is open source software with self-published libraries that are not open source. A binary file in FT's source repository is not sufficient to claim FT uses and created it, but is keeping the source closed. If this happens to be the case, "mostly open source software" is a poor construction; "FT is open source software; however, it references necessary libraries that are closed source," is more accurate and more clear. 100DashSix (talk) 15:45, 16 July 2009 (UTC)


 * Since the noted dcu files are in the FT source depot and cannot be found any other place on the internet, the burden of proof rests on proving they were *not* developed by the FT developers for use with FT. However, here is proof anyway. Unless you can provide sources indicating otherwise, the text will be returned to indicating that FreeTrack is only partially open-source.


 * The statement "FT is open source software" asserts that the whole FreeTrack project is entirely open source and it takes only one single counter example to invalidate the statement. The so-called "necessary libraries" are part of FreeTrack, not externalities. Whereas the following is an accurate, valid and supportable fact : "FreeTrack is free, mostly open source software, however its implementation of the TrackIR interface is only provided in compiled binary format with no source available to the public".


 * TIRServer.dcu and TIRTypes.dcu are the only pre-compiled delphi binaries in the FT 2.2 depot


 * TIRServer.dcu and TIRTypes.dcu cannot be found as of this date on any other site in the internet than the FT depot when searching with google


 * Both files contain numerous strings bearing the name "freetrack" as well as ones indicating use for TrackIR compatibility. Such as "FREETRACKSHAREDMEMORYAREA", "FREETRACK_MUTEX", "TIRPROVIDER", "NP_CLIENT_LOCATION", "TIRViews", "TRACKIRDATA"


 * The source for TIRTypes.dcu was available as TirTypes.pas in previous versions of the FT (V0.2 tag in the depot for example). The TirTypes.pas source contains the same such strings as "FREETRACKSHAREDMEMORYAREA = 'FreetrackSharedMem';", "TRACKIRDATA = packed record". As noted earlier, the source has been removed in current versions.


 * The TIRServer.dcu file references TIRTypes : "TIRTypes"


 * Other source files in the FT 2.2 depot reference TIRServer and TIRTypes. PoseDataOutput_fm.pas : "SimConnect_dm, FPCUser, ProfilesMngr_fm, FTServer, TIRServer, TIRTypes;" —Preceding unsigned comment added by 75.145.66.140 (talk)


 * TIRServer.dcu and TIRTypes.dcu being in the FreeTrack repository is not sufficient to say TIRServer.dcu (or indeed any other file) is necessary to compile FreeTrack, nor enough to say it violates the GPL. A valid citation would be documentation or forum posts showing how to compile FreeTrack, and/or discussing the files in question. All the citation says at the moment is "as of (date), the files (files) are only provided in the FreeTrack repository in a binary form." As for FreeTrack being open source, there are numerous sources that support that claim.


 * If FreeTrack claims it's open source, but there is published information saying otherwise, then it's perfectly valid to call it "mostly open source," or "open source, BUT...," or "claims it's open source but violates the GPL." All your claims, however, are original research. Running 'strings' on the binaries and noting that "trackIR" and "freetrack" comes up is not a published source. The burden of proof is on Wikipedia editors to support their claims with valid citations and published research. 100DashSix (talk) 19:34, 21 July 2009 (UTC)

TrackIR cleanup Feb 2010
The interface and controversy sections had significant overlap so I merged and tightened them. Since I can't find any reputable third party publications discussing controversy regarding this issue I stuck to the more neutral 'interface' title instead of editorializing by making claims of controversy.

Added a OR and Ref template.

Require official citations for NaturalPoint stating that FreeTrack violates their copyright and the FreeTrack statements in response.

The string paragraph is all original research, the specifics are not particularly notable and difficult to verify due to it being closed source. Need a secondary source to warrant going into so much detail. Without secondary source, can only really state accusation of copyright violation from a primary source.

TIRServer.dcu commentary is original research, very specific and not notable since it is obvious that the closed TrackIR interface cannot be made open source by a third party. Also it repeats the same point in the prior section, that the interface is closed source. Open source software using a closed source interface library that comprises a small percentage of the total code is not controversial or unusual.

Quoting the TrackIR software EULA is unnecessary, the circumstances make violation obvious, only need to state that there is a violation and reference where to find the license.

Updated information on the TrackIRFixer.

Hplusplus (talk) 15:41, 23 February 2010 (UTC)

Regarding 75.145.66.140 reverting my contributions:

It is unnecessary and possibly incorrect to claim there are three TrackIR interfaces. Seem more like different versions.

'Most' new games using encryption is a more accurate description than 'some', there are 5 (see TrackIRFixer) that do and 2 (Rise of Flight and Birds of Prey) that don't. Those that don't probably received the SDK before encryption was introduced.

TrackIRFixer name is more precise than 'patch', first reference is not TrackIRFixer primary source but insignificant secondary source for games that don't use encryption. Second reference is a secondary source containing opinionated discussion of TrackIRFixer with little credibility, unknown SimHQ identity claims 'I think that NaturalPoint have encrypted their API '. The third reference regarding backward compatibility is on the NaturalPoint forum and is 'Access Denied'. The final reference is the primary source for TrackIRFixer and is incorrectly named 'Freetrack title patching'. TrackIRFixer is for TrackIR games and not specifically for FreeTrack. The TrackIRFixer primary source is important for establishing the timeline, the existence of encryption and the existence of the 'patch' itself and should be referenced accordingly.

The incompatibility with some DRM is notable and logical. Anti-cheat incompatibility is only for some, not all, anti-cheat. Since it doesn't affect any present games TrackIRFixer works with it has limited notability.

Previous description of the TrackIR interface, requiring signed license agreement and NDA, is a more accurate description and has a reliable primary source and credible secondary source. It is also important that this introduces the section so the rest can be put into context.

TIRViews function is notable and logical, uses interfaces like SimConnect where available and memory hacks otherwise (only way).

EULA instantiation should be avoided, prior description more concise and neutral. Given that there is a violation, it is obvious that TIRViews is only available under EULA, obvious that it must be loaded, and obvious that the EULA opposes this. This is all normal EULA operation that is explained in EULA page. Another reason for not putting undue weight into TIRViews is that it provides additional support to only 5 of the 103 TrackIR Enhanced games available, the most recent of which is 6 years old.

The 'List of games supporting FreeTrack' is misleading, the only game 'supporting FreeTrack' is ARMA2.

'See also' section is a "A bulleted list, preferably alphabetized, of internal links (wikilinks) to related Wikipedia articles"

Hplusplus (talk) 17:08, 13 March 2010 (UTC)

In-device pre-processing
The in-device pre-processing column in the camera comparison table is ambiguous and not factual. Many webcams have in-device pre-processing in the form of video compression, for example the PS3 Eye has on-chip JPEG compression which significantly improves cpu efficiency. TrackIR's in-device pre-processing is not altogether different, being an extreme form of compression. The method of processing is made clear in the output column and the benefit is clear in the cpu usage column, so the in-device pre-processing column is an inferior and redundant representation. Seemoss 18:49, 11 May 2010 (UTC)

Subpixel precision comparison
Any digital image can have blob centroids extracted down to ~0.06 pixels for binary and ~0.007 pixels for grayscale. This is the product of interpolation and is variable, based on the method of analysis and the blob size and shape. Except for the Wii remote which does its own centroid calculations (poorly), it is not a discrete fixed property of a camera. For this reason video cameras do not advertise subpixel precision, but since they produce digital grayscale images, we know they have more potential subpixel precision than binary TrackIR and the exact same as grayscale TrackIR. This needs to be clearly shown in the table.

Since TrackIR provides a specific subpixel precision value for grayscale, there is little option but for the other grayscale cameras to adopt this same precision value in the table, to emphasize the direct relationship between output format and precision. The exact value is fairly arbitrary, being used more for its symbolic value to permit comparison. Remember this is a camera comparison table only and reflects the potential camera capabilities independent of software. Seemoss 18:16, 12 May 2010 (UTC)

Request for plug - reversed game protocol/tir5 'enhanced' support
Can someone mention opentrack in here? Personally can't, due to conflict of interest. 2001:470:600D:DEAD:BEAE:C5FF:FEE1:4407 (talk) 11:31, 23 October 2013 (UTC)

Please at least mention that both sides of protocol are reversed by "us": game protocol w/o NDA, as well as TIR5... 2001:470:600D:DEAD:BEAE:C5FF:FEE1:4407 (talk) 11:32, 23 October 2013 (UTC)