Filetype for RAW image files

Is there yet a filetype for RAW data created by digital cameras or scanner?
I found a filetype ‘RAW’, but
http://en.wikipedia.org/wiki/List_of_RISC_OS_filetypes
says FF1 RawData Raw unprocessed data (e.g.terminal streams)

Is one filetype for raw image files suitable?
In the windows word there are a lot of extensions, e.g.
.3fr, .ari, .arw, .bay, .crw, .cr2, .cap, .dcs, .dcr, .dng, .drf,.eip, .erf, .fff, .iiq, .k25, .kdc, .mdc, .mef, .mos, .mrw, .nef,.nrw, .obm, .orf, .pef, .ptx, .pxn, .r3d, .raf, .raw, .rwl, .rw2,.rwz, .sr2, .srf, .srw, .×3f
from wikipedia.
Even with the same extension, the internals may be different.So a selfrespecting editior must not trust the extionsion.
And one filetype may be enough in the RISC OS world

I think a new filetype should be in the range E00 – FFF,
as this a generic data type.

The textual equivalent could be ‘RAWImage’

Any opinions?

http://en.wikipedia.org/wiki/List_of_RISC_OS_filetypes

Wow. That page needs a lot of love.

Even with the same extension, the internals may be different.So a selfrespecting editior must not trust the extionsion.
And one filetype may be enough in the RISC OS world

There are two ways to look at this. The first is to lump anything that is some sort of “raw” image into one filetype and let the conversion software sort out the mess. The alternative is to have different filetypes for different sorts of raw file (just as BMPs and PNGs and such are all image formats but have different filetypes).
My opinion is the former – one filetype and let the convertor deal with it. This is because for most people, they won’t come across raw that often so there is not so much call for claiming many filetypes for something specialist.

says FF1 RawData Raw unprocessed data (e.g.terminal streams)

I use FF1 myself for dumps from, say, serial ports or parallel ports. Data read and stashed to be processed later gets an &FF1 file. While a raw image from a camera could well count, you would need to make processing software capable of handling all sorts of things (including coping with single-byte files!¹).

I think a new filetype should be in the range E00 – FFF, as this a generic data type.

But a specialist one. On the other hand, is GIF (&695) not generic? How about JPEG (&C85)? Or Zip (&DDC)? Or CSV (&DFE)?

That said, looking at the list I find myself asking WTF Acorn were smoking. Commadore Amiga? Unix executable? What? Sure, filetypes, yes. But not in the generic section! And PC Emulator Configuration is just taking the p……

¹ I’ve known some software to crash if the file is not empty, but too small for it to be able to read the expected header… Thankfully for use, I’m talking Windows, not RISC OS, but it is something to take into consideration.

On the other hand, is GIF (&695) not generic? How about JPEG (&C85)? Or Zip (&DDC)? Or CSV (&DFE)?

That depends on who allocated them, surely? If (say) &DDC was allocated for (say) SparkFS, it would presumably go into the third-party area and not Acorn’s.

That said, looking at the list I find myself asking WTF Acorn were smoking. Commadore Amiga? Unix executable? What? Sure, filetypes, yes. But not in the generic section! And PC Emulator Configuration is just taking the p…

And that might depend on age. I have a dim memory (on which I’m willing to be corrected) that pre-reshuffle &E00-&FFF was “reserved for Acorn” and &A00-&AFF was “reserved for Acornsoft”. A lot of those types are likely to have been allocated in the days before Acorn realised that 4096 types wasn’t that many.

Right back at the start, I suspect things were even less regimented. Things like &BBC are a bit tricky to explain otherwise…

needs a lot of love.

That’s why it says This list is incomplete; you can help by expanding it. :-P

asking WTF Acorn were smoking

Yes, many of the commercial image types are now considered generic, but I guess this may not have been obvious at the time.

&BBC

I love that one!

Is there yet a filetype for RAW data created by digital cameras or scanner?

Returning to the original question, perhaps the answer is no. Does any image manipulation software accept raw camera data? I’d be inclined to use &FFD and have the software determine the type (assuming it’s simple enough to decode) from the within the file. I’ve not personally come across a generic raw data format for such images, with each manufacturer apparently using their own proprietary format.

That’s why it says This list is incomplete; you can help by expanding it. :-P

I could paste in some of the filetypes I am aware of – but there is no published reference for some of that stuff ¹; is that acceptable or is somebody eventually going to [citation needed] me?

¹ …and in many cases no published reference that I can be bothered to look for, I have little enough free time as it is. To be honest, if a file type of &XXX loads and does something on RISC OS, I would consider that adequate evidence that &XXX is SuchAndSuch a file, but bl**dy Wiki might consider that to be [WP:NOR] (where being a stickler for rules is more important than accuracy).

I’ve dropped some into the Talk page: http://en.wikipedia.org/wiki/Talk:List_of_RISC_OS_filetypes

Are we all aware of this one? It would seem to be a good place to start.

It’s difficult after all this time to trace the provenance, but I’d add to that list:

I’ve been using these for a long time though some seem to clash:

Also FC9 WiniUtil and FD5 RiscOs3, and many of the filetype names, are wrong.

I’m not going to contribute to the Wikipedia page. I’d prefer the information was in some other Wiki that Wikipedia referenced or mirrored.

I’m not going to contribute to the Wikipedia page. I’d prefer the information was in some other Wiki that Wikipedia referenced or mirrored.

Like this one?
http://www.riscosopen.org/wiki/documentation/show/File%20Types

Oh well, in that case, time to edit!

Thanks, Rick :-)

[Forgot to refresh my browser.]

I’d prefer the information was in some other Wiki that Wikipedia referenced or mirrored.

That’d be fair enough, but WP:SELFPUBLISH doesn’t regard open wikis as reliable.

perhaps the answer is no

Maybe the RISC OS port of dcraw has some info (there’s no obvious mention of RO filetypes in its documentation).

OK, I’ve simplified the markup quite a bit, but it still requires editing in a separate application. I propose making a number of changes:

Remove Category column

However noble the original intention, this categorisation wasn’t reflected in actual usage and is now utterly confusing. What, in any case, does “commercial” mean in relation to MPEG for example?

Remove lists of applications

Explaining which program generates a proprietary filetype is useful, but listing every program that loads text files is not.

Add allocation status

It would be useful to know if a filetype has been officially allocated, or has simply been found in the wild.

That’d be fair enough, but WP:SELFPUBLISH doesn’t regard open wikis as reliable.

I couldn’t give a monkey’s what Wikipedia thinks. And that’s putting it mildly.

Some more ‘found in the wild’:
6B1 DataFile LTGame
6B2 ArryFile LTGame
6B3 JnyFile LTGame
6B4 TTFile LTGame
EFB BasicChn (to -chain rather than -quit a BASIC programme)
9EB Index (not used any more)
9F0 Money !Money

all of which I just used without bothering to register them.

Also

132 ICO (imagemaster aka DPlgScan)
3d6 (stronghelp)
A7E MSWordX (techwriter)
A81 ODT (techwriter)
aba CPIO (sparkfs)
abf Cabinet (sparkfs)
B24 OVPScrpt (ovation pro)
B25 OvPDDL (ovation pro)
B26 OvPStyle (ovation pro)
B27 OvnPro (ovation pro)
b56 BubStyle (bubblehelp)
df6 CDImage (cdromfs)
e1f (elfrun)
f81 JSScript (phoenix)
f91 URI (pheonix)
f98 PhotoShp (photodesk)

1D9 Syncjob, used by !SyncDiscs, officially registered.

Officially registered:
155, Filelist, DigitalCD playlist
1A8, Vorbis, Ogg Vorbis audio files in Ogg container, .ogg .oga, audio/ogg audio/x-ogg audio/x-vorbis audio/x-vorbis+ogg application/ogg application/x-ogg
1AD, AMPEG, you can add audio/mp2 audio/mp3 audio/x-mp3 to the existing list
1AE, ImpulsTk, ImpulseTracker soundtracks, .it
1AF, FstTk, FastTracker2 soundtracks, .xm
1B0, ScreamTk, ScreamTracker soundtracks, .s3m .stm
1B1, PCTk, generic file type for PC soundtracks, .afm .dbm .digi .dsm .dtm .far .mtm .ptm .ult .669
1CF, FLAC, Free Lossless Audio Codec, .flac, audio/flac audio/x-flac
A8D, VOB, MPEG DVD-video files, .vob, video/x-vob
A8F, AC3Audio, AC3 raw audio, .ac3, audio/ac3
AA7, M3UFile, mp3 URL list (WinAmp), .m3u .vlc, audio/mpegurl audio/x-mpegurl application/m3u
AA8, pls, internet radio playlist (SHOUTcast), .pls, audio/x-scpls application/pls
F80, XML, eXtended Markup Language files, .xml, /xml */+xml

Trevor – further to my Talk page post, I can add that I use &112 as a Teletext file in my !Teletext. I used this as it was the same filetype/format as Ground Control’s !TTX. I assumed they would have registered it, but you’ll need to confirm with The Big List Owner. [I thought !TTX was crap so I rolled my own ;-) ]

VideoList file, &17F, is my own, it is a registered type. [source]

I would support removing the Commercial/Noncommercial distinction. This cannot, AFAIK, be maintained as I’m not aware that !Allocate even asks this. Plus, there are several public allocations in the Acorn Reserved area, as I noted in the Wiki talk page (PDF (&ADF) and MSWord (&AE6)); thus maintaining this anachronism will just be confusing. It’s how it was intended, sure, but not how it eventually ended up.

I would especially support a reference to “known to be officially allocated”; for these are real filetypes. Anything unallocated is just winging it…

Interesting. I just tried setting a file on my Iyonix here to F80. The icon changed to a proper XML one, but the filetype remains described as &F80, so there isn’t a textual name for it.

Which makes me ask the question: where are all these types assigned? Are they all in the MimeMap file?

And how come my Iyonix has half the information?

That’d be fair enough, but WP:SELFPUBLISH doesn’t regard open wikis as reliable.

Wiki saying that open wikis can’t be considered reliable.

The irony is killing me.

You might have guessed by nemo‘s reply that some editor policies have p!$$ed people off.
One of the things that concerns me is we might make a nice list of filetypes for some jockstrap butthurt <the rest is censored> to come along and decide it isn’t encylopaedic, don’t cite enough references, and nobody except us few gives a crap about such trivia; nominated for deletion…..
For a project that wants to collect together the world’s knowledge, Wiki seems to have some damned strange ideas. Not to mention editors who have the opinion: “I own this topic, mine! Mine! MINE! My precious!”.

While it is rather astonishing to see a list of RISC OS filetypes on Wikipedia, perhaps the most important question I could ask is: Why isn’t it here?

On ROOL’s site is the logical place to catalogue filetypes; and I’m fairly certain that ROOL wouldn’t get hung up over discovery of filetype use by looking at !Boot files (aargh! no citation! aargh! original research! aargh!) because having a coherent list is the important thing.

There is the start of a list (and quite a long list) here:
https://www.riscosopen.org/wiki/documentation/show/File%20Types

I would like to make two suggestions:

1. Delete the left hand column. It’s annoying to see “Personal use only” (etc) repeated a billion times. The list should be split into three sections:
   Personal use filetypes (&000 → &0FF), Registered filetypes (commercial and non-commercial; &100 → &9FF, &B00 → &DFF)), Operating System/Acorn filetypes (&A00 → &AFF, &E00 → &FFF). It looks like the list will be five sections due to Acorn giving themselves all the A** types (oh, very droll…)). If we call everything else just “Software”, we can get around “Generic” not containing such utterly generic things as CSV files, but instead having such ludicrous things as Sun raster images; but then the filetype listing is chock full of eccentric quirks!
2. As before, an asterisk or something on a new right-hand column if the filetype has been officially registered.

Thoughts?

Who cares what category they are in. Get rid of the first column and get rid of the & in front of the filetypes.

Thoughts?

My first thought was you didn’t read all the postings in this thread…

Other than that, I agree with the changes nemo suggested.

It would be useful if ROOL would release a listing of all official filetype allocations. However, such a listing was always treated as ‘commercially sensitive’ and therefore confidential. This cannot be the case now.

Which makes me ask the question: where are all these types assigned? Are they all in the MimeMap file?

Is it not the app owning the filetype that sets the textual name in its !Boot or !Run file. Eg.

Set Alias$@RunType_1D9 /<SyncDiscs$Dir>.!Run %%*0
Set File$Type_1D9 Syncjob