On Friday 06 February 2009 00:00:21 Ludovic Rousseau wrote: > > pcsc-lite has some infrastructural limitations, that cannot be > > easily fixed. it has no driver support environment, so > > writing new drivers is difficult. > > Do you need a new driver for pcsc-lite? For which reader? > Do you know a smart card reader that is not CCID these days?
If there is none, why do we need a complex framework? Drop the IFD Handler API from pcsc-lite and make it much simper... > > Ludovic made it clear that he does not want to cooperate. So I guess > > the initiative is rejected and dropped. > > The problem is that I do not have an unlimited amount of time. I > prefer to slowly improve pcsc-lite step-by-step rather than start a > new project from scratch. Which, as I estimate, will take more resources and never get you to a point it is feature complete and maintainable. And for the 15 or so time, I took upon my-self most of the work... Forget it. Sorry, but nothing you wrote convinced me that I am wrong. > > I have enough experience to know who many resources should be > > invested in pcsc-lite in order to make it right. And it is more than the > > resources required for a rewrite. > > It depends on your definition of "make it right". pcsc-lite already I tried to explain, there is much more in that. The suspend/resume issue for example, which is undebugable in current state, maybe I am the only one who actually tried to use the software for real. > works for many people. So is DOS... I remember the same kind of arguments when people migrated to Windows.... Vendors continued to support DOS applications using the same argument "It works under Windows too". Whenever Windows user had an option they switch to a better solution. When Windows NT came up, Microsoft made extreme effort of backward compatibility, but many of these companies that survived Windows 3.1, disappeared. BTW: I can still run Digger on my Linux machine. I regret I could not help more changing the reality. > Yes, it can be improved to be more CPU efficient. It is on my todo > list. My idea is to: > - start pcscd only if/when an application calls SCardEstablishContext() How exactly will you start the daemon? I guess you going to add another dependency, for example dbus. Or have a setuid wrapper... Which is not a wise to do these days. Also handling the race condition of multiple instances is an issue. > - start a reader driver only when an application wants to use this > reader (and not just when the reader is connected as you propose). Whatever. > Help is welcome. In your terms, of course... > > I am considering, just for fun, to add slot event reporting into OpenCT > > and add PC/SC compliant interface that uses OpenCT API. It is not the > > framework I outlined, but it will take us much closer to what we need. > > Example: We have daemonless standalone PC/SC implementation. > > You want to start a competing PC/SC library? Do you think that will > simplify the live of normal users to have two PC/SC frameworks? As I said, I think you are wrong, and continue to mislead users. And as I wrote above, it is just for fun for the time being. > Maybe you should improve OpenSC to make it stateless instead. THAT > would help every OpenSC users. Sure. But there is also PKCS#11 slot events issue to solve. And in order to make this simpler, I need to reduce the number of reader frameworks OpenSC supports. And as I am going to continue to use OpenCT, and understand the need of PC/SC I think this is the way to go. Also, an implementation of PC/SC in client library, without server part will help in the discussion that will take place about two years from now. As if only CCID driver will remain, it would be very simple to use it with a simple reader driver only implementation. > Or add/improve support of (nearly) deployed cards. The IAS ECC > (Identification Authentication Signature - European Citizen Card) card > should be deployed in many European countries in 2010. The same card > specification will be used for health cards. > Search for "IAS ECC" or "CEN/TS 15480". [1] is a (now old) description > of the project. > > If nobody invests in OpenSC internals knowledge the new cards will not > be supported and European citizens will not have free software > solution to use their cards. I do not live en Europe. > pcsc-lite is (well) maintained. OpenSC is on the decline. You should > know what to work on if you want to help support of smart cards in > free software project. It depends of what you perceive as maintained. Thank you for discussing this. Alon. _______________________________________________ opensc-devel mailing list opensc-devel@lists.opensc-project.org http://www.opensc-project.org/mailman/listinfo/opensc-devel
