Hey there, saw a post at Classicgaming.com asking us to give our thoughts.

I think it is a worthwhile idea, rather like a classicgaming version of Xfire, but only if there is an option to login as an anonymous user. I doubt anybody wants to be served a subpoena to hand over their information on users, for example.

Second, I think any server that is accepting requests worldwide is going to be a magnet for hackers as well as put under a lot of strain normally, though secure sessions would help mitigate that.

Finally, I think the debate about identifying games is wasteful. This would be a nice way to track how many different versions of ROM and disk images there are floating about the internet; allowing users to select the name (or submit one for checking if it's not found in the list, to prevent misspellings) plus any known status (i.e. the game includes a trainer or an intro hack) if their image's CRC doesn't match up.

Manuel calls the GPL to the rescue:

Quote:

Quote: Now would it be spyware? [...] Such modified versions would qualify as spyware. Yes, but such software would be open source, so you could verify this if you don't trust it. Or, just don't run these versions if you don't trust them, run the official distribution. Thank you, GPL.

Yes, I could do that, in principle, but if the program contains GameStats, it could be enabled anywhere in the program by a carefully crafted bug. So if want to verify that a program with GameStats is not spyware, I would have to read and verify each and every file in the source code.
If a program doesn't contain GameStats, I would have far less to worry about.

(If I have any concern that the binary may not match the sources I would recompile as well, but such as mismatch would be in violation of the GPL.)

Hmm, another highly unlikely event is used as a counter-argument not to implement GameStats. You can what-if until infinity with reasoning like that. Point is, not only is such a bug unlikely, beta-testers and (in an already less likely event) end-users will probably spot such a bug in no-time due to their firewall yelling that their emulator is trying to send data over the web. Especially now that we know mars2000you will be watching over our backs ^_^. Sure, there would be less to worry about this if the emulator would not contain gamestats, but then again... you would never develop any feature with a "suppose it contains a bug"-approach like this. What about the much desired multiplayer gaming over the web? Or emulation of Obsonet, or...

Yes, the openMSX (e.g.) authors could have put in the feature by a carefully crafted bug. You'd have to trust us that we won't do that.
If you are talking about other people's versions of the same program: again it's a matter of trust. If you don't trust it, run only the offical version. If you don't trust the authors of the official version: don't run the program.

If the gamestats would be implemented e.g. in TCL (likely in case of openMSX), you could just erase the script and feel happier, maybe!

Oh, and snout: most firewalls don't alert you when a program sends data over the web, if they use the normal web port (HTTP, 80).

snout: good point! There will be more internet-related features in the future, probably. And those will not work via port 80, I guess...

Quote:

EdOscuro wrote: Finally, I think the debate about identifying games is wasteful. This would be a nice way to track how many different versions of ROM and disk images there are floating about the internet; allowing users to select the name (or submit one for checking if it's not found in the list, to prevent misspellings) plus any known status (i.e. the game includes a trainer or an intro hack) if their image's CRC doesn't match up.

Good point. I was thinking of such an approach as well for disks while reading the thread
Even though it could be abused by people by submitting a different name for a different game, it would still be a nice alternative.
Maybe have those entries marked with an 'unconfirmed' flag or something. Or perhaps have those combined with a screenshot at a certain point upon submission as verification of the game.

Although the idea can be fun and handy in a way (especially to see the GenMSX database expanded with more screenshots and such), I'd still would opt for the GameStats being a separate module that I could choose not to install.

Not because of fear of privacy/security-breaches, but because I'd like to have emulators as bloat free as possible..
If I don't use a feature, it would be nice if I didn't have to have it in the program either..

Quote:

Manuel wrote: If the gamestats would be implemented e.g. in TCL (likely in case of openMSX), you could just erase the script and feel happier, maybe!

That sounds like a nice solution

As for the replay mode, could it also be used as some form of netplay as well?
sharing the replayinfo realtime to allow it being displayed at more locations over the net so people could spectate or even participate?
*thinks of playing bombaman via the net through it *
or would that be 'impossible' due to lag for instance?

Fixato, indeed, the same (kind of) data is needed to sync emulators over the network, for a netplay thing. So this kind of functionality enables a lot more interesting features than just replay!

how about putting the info on that feature in a separate topic? Since it has a wider scope than this gamestats topic.

Well, it's not in my power to do that, but it's a bit intermingled now with this topic indeed.
In any case, dvik started to work on the event recording, but I started to work on a proof-of-concept implementation as well (in openMSX, of course). So, we'll get to hear a lot more about this in the (near) future anyway.

(time for a bit of fun intermission. The first MSX related spyware? think again!)

Quote:

Finally, I think the debate about identifying games is wasteful. This would be a nice way to track how many different versions of ROM and disk images there are floating about the internet; allowing users to select the name (or submit one for checking if it's not found in the list, to prevent misspellings) plus any known status (i.e. the game includes a trainer or an intro hack) if their image's CRC doesn't match up.

If you're talking about a CRC, aren't you exactly doing what you are stating to be against? Identifying games? (To check if it's not been tampered with...)

By the way, what is the problem in knowing how many versions are floating around the net? There already exist lists like that (I forgot the name) for many computer systems.

Quote:

(time for a bit of fun intermission. The first MSX related spyware? think again!)

Well, this kind of spyware must made by the devil himself, last week I had the same kind of spyware on my computer, only a different name, "virtumonde" in this case.
The only thing I could do, was format the hd and install windows again

Quote:

(time for a bit of fun intermission. The first MSX related spyware? think again!)

Also for the fun - do you remember these news ?

Now check this site !

What's in a name, eh?