Post by thread:[EE]: Internet devices + java timing

Hi, Problem: Internet enabling of a device purely having serial comms via the internet. Suppose that a simple device purely supports RS232 and has a modem so it may be dialled up. Let's say that such a device operates quite a strict timing regime with respect to message turnaround times and such like which could cause some timing issues if serial comms is not handled to an accurate timebase/timeframe. To Internet enable this same device could, at a simplistic level, require the following: 1. Browser running java applet which can talk to a central java application via a socket 2. Central, muti-threaded java application servicing applet clients. This app also perform multi-threaded serial comms to a modem....which is attached to the RS232 device. 3. RS232 device with modem. This is in the configuration of: < Internet > < Server side Device side modem link > Applet<----TCP/IP----->Server app<--serial-->modem<---------->modem <---serial---->serial device In order to contact the serial device, the applet tells the server what to dial, then the connection is initiated via modem. The server/modem can then hold the telephone line to the RS232 device while data is transferred between the system end points. When the user presses the "end session" button on the applet, the server drops the modem connection and the.....session ends, predictably. Of course, this scenario could be handled in CGI/Perl but let's leave it at Java for the moment. I hear it is tricky to perform timing-sensitive serial comms in Java due to the nature of the beast. To use another app to talk serial (i.e. one written in C/C++) is, I imagine, a no-no for most ISPs as would be calling C functions from Java (if that is possible). Has anyone ever implemented such a service? If what I heard about the timing issues is actually correct, any comments on how it is possible to force Java to behave itself serially/interface Java to something that can perform serial comms with an accurate timebase? Please forgive my ignorance on the subject. Regards, Dan -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

To me this sounds like something for the TINI - doesn't it? http://www.junun.org/TINI/ Sven in Sweden -----Ursprungligt meddelande----- Från: D Lloyd <spam_OUTdan.lloydTakeThisOuTGB.ABB.COM> Till: .....PICLISTKILLspam@spam@MITVMA.MIT.EDU <PICLISTKILLspamMITVMA.MIT.EDU> Datum: den 9 oktober 2001 18:43 Ämne: [EE]: Internet devices + java timing {Quote hidden}>Hi, > >Problem: Internet enabling of a device purely having serial comms via the >internet. > >Suppose that a simple device purely supports RS232 and has a modem so it >may be dialled up. Let's say that such a device operates quite a strict >timing regime with respect to message turnaround times and such like which >could cause some timing issues if serial comms is not handled to an >accurate timebase/timeframe. > >To Internet enable this same device could, at a simplistic level, require >the following: > >1. Browser running java applet which can talk to a central java application >via a socket >2. Central, muti-threaded java application servicing applet clients. This >app also perform multi-threaded serial comms to a modem....which is >attached to the RS232 device. >3. RS232 device with modem. > >This is in the configuration of: >< Internet > < Server side >Device side modem link > >Applet<----TCP/IP----->Server app<--serial-->modem<---------->modem ><---serial---->serial device > >In order to contact the serial device, the applet tells the server what to >dial, then the connection is initiated via modem. The server/modem can then >hold the telephone line to the RS232 device while data is transferred >between the system end points. When the user presses the "end session" >button on the applet, the server drops the modem connection and >the.....session ends, predictably. Of course, this scenario could be >handled in CGI/Perl but let's leave it at Java for the moment. > >I hear it is tricky to perform timing-sensitive serial comms in Java due to >the nature of the beast. To use another app to talk serial (i.e. one >written in C/C++) is, I imagine, a no-no for most ISPs as would be calling >C functions from Java (if that is possible). > >Has anyone ever implemented such a service? If what I heard about the >timing issues is actually correct, any comments on how it is possible to >force Java to behave itself serially/interface Java to something that can >perform serial comms with an accurate timebase? Please forgive my ignorance >on the subject. > >Regards, >Dan > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

Hi, "For the Tini" in what respect? I was thinking of a solution without having to deploy any TCP/IP devices at the embedded appliance end of the system; the embedded system would preferably retain just a RS232 capability with just a plain old modem. This means no extra hardware costs or firmware changes, too. To my mind, it makes far more sense to use some complicated software in a central place (probably called through or in cooperation with a standard web server running on a very capable platform) than increasing the hardware end costs of the embedded systems. If the embedded devices have access to a network (like in a factory), then that is another story. Regards, Dan To: .....PICLISTKILLspam.....MITVMA.MIT.EDU cc: From: "Milton Medicinteknik KB, Vikingstad, Sweden" <EraseMEmilton.medicinteknikspam_OUTTakeThisOuTTELIA.COM> Subject: [EE]: Internet devices + java timing To me this sounds like something for the TINI - doesn't it? http://www.junun.org/TINI/ Sven in Sweden -----Ursprungligt meddelande----- Från: D Lloyd <dan.lloydspam_OUTGB.ABB.COM> Till: @spam@PICLISTKILLspamMITVMA.MIT.EDU <KILLspamPICLISTKILLspamMITVMA.MIT.EDU> Datum: den 9 oktober 2001 18:43 Ämne: [EE]: Internet devices + java timing {Quote hidden}>Hi, > >Problem: Internet enabling of a device purely having serial comms via the >internet. > >Suppose that a simple device purely supports RS232 and has a modem so it >may be dialled up. Let's say that such a device operates quite a strict >timing regime with respect to message turnaround times and such like which >could cause some timing issues if serial comms is not handled to an >accurate timebase/timeframe. > >To Internet enable this same device could, at a simplistic level, require >the following: > >1. Browser running java applet which can talk to a central java application {Quote hidden}>via a socket >2. Central, muti-threaded java application servicing applet clients. This >app also perform multi-threaded serial comms to a modem....which is >attached to the RS232 device. >3. RS232 device with modem. > >This is in the configuration of: >< Internet > < Server side >Device side modem link > >Applet<----TCP/IP----->Server app<--serial-->modem<---------->modem ><---serial---->serial device > >In order to contact the serial device, the applet tells the server what to >dial, then the connection is initiated via modem. The server/modem can then >hold the telephone line to the RS232 device while data is transferred >between the system end points. When the user presses the "end session" >button on the applet, the server drops the modem connection and >the.....session ends, predictably. Of course, this scenario could be >handled in CGI/Perl but let's leave it at Java for the moment. > >I hear it is tricky to perform timing-sensitive serial comms in Java due to >the nature of the beast. To use another app to talk serial (i.e. one >written in C/C++) is, I imagine, a no-no for most ISPs as would be calling >C functions from Java (if that is possible). > >Has anyone ever implemented such a service? If what I heard about the >timing issues is actually correct, any comments on how it is possible to >force Java to behave itself serially/interface Java to something that can >perform serial comms with an accurate timebase? Please forgive my ignorance {Quote hidden}>on the subject. > >Regards, >Dan > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Hmmmm.. I agree with you. I am involved with a project at the moment where data is collected with PIC's, and then transferred with a TINI. My idéa was to do approximately what you suggest, but we are many involved and the concept IS TCP/IP in this case. BUT, basically, it is all - of course - very case-to-case dependant. Sven in Sweden. -----Ursprungligt meddelande----- Från: D Lloyd <RemoveMEdan.lloydTakeThisOuTGB.ABB.COM> Till: spamBeGonePICLISTspamBeGoneMITVMA.MIT.EDU <TakeThisOuTPICLISTEraseMEspam_OUTMITVMA.MIT.EDU> Datum: den 10 oktober 2001 08:46 Ämne: Re: [EE]: Internet devices + java timing {Quote hidden}>Hi, > >"For the Tini" in what respect? I was thinking of a solution without having >to deploy any TCP/IP devices at the embedded appliance end of the system; >the embedded system would preferably retain just a RS232 capability with >just a plain old modem. This means no extra hardware costs or firmware >changes, too. > >To my mind, it makes far more sense to use some complicated software in a >central place (probably called through or in cooperation with a standard >web server running on a very capable platform) than increasing the hardware >end costs of the embedded systems. If the embedded devices have access to a >network (like in a factory), then that is another story. > >Regards, >Dan > > >To: RemoveMEPICLISTTakeThisOuTMITVMA.MIT.EDU >cc: >From: "Milton Medicinteknik KB, Vikingstad, Sweden" ><milton.medicinteknikEraseME.....TELIA.COM> >Subject: [EE]: Internet devices + java timing > > > >To me this sounds like something for the TINI - doesn't it? > >http://www.junun.org/TINI/ > >Sven in Sweden > > >-----Ursprungligt meddelande----- >Från: D Lloyd <EraseMEdan.lloydGB.ABB.COM> >Till: RemoveMEPICLISTEraseMEEraseMEMITVMA.MIT.EDU <RemoveMEPICLISTspam_OUTKILLspamMITVMA.MIT.EDU> >Datum: den 9 oktober 2001 18:43 >Ämne: [EE]: Internet devices + java timing > > >>Hi, >> >>Problem: Internet enabling of a device purely having serial comms via the >>internet. >> >>Suppose that a simple device purely supports RS232 and has a modem so it >>may be dialled up. Let's say that such a device operates quite a strict >>timing regime with respect to message turnaround times and such like which >>could cause some timing issues if serial comms is not handled to an >>accurate timebase/timeframe. >> >>To Internet enable this same device could, at a simplistic level, require >>the following: >> >>1. Browser running java applet which can talk to a central java >application >>via a socket >>2. Central, muti-threaded java application servicing applet clients. This >>app also perform multi-threaded serial comms to a modem....which is >>attached to the RS232 device. >>3. RS232 device with modem. >> >>This is in the configuration of: >>< Internet > < Server side >>Device side modem link > >>Applet<----TCP/IP----->Server app<--serial-->modem<---------->modem >><---serial---->serial device >> >>In order to contact the serial device, the applet tells the server what to >>dial, then the connection is initiated via modem. The server/modem can >then >>hold the telephone line to the RS232 device while data is transferred >>between the system end points. When the user presses the "end session" >>button on the applet, the server drops the modem connection and >>the.....session ends, predictably. Of course, this scenario could be >>handled in CGI/Perl but let's leave it at Java for the moment. >> >>I hear it is tricky to perform timing-sensitive serial comms in Java due >to >>the nature of the beast. To use another app to talk serial (i.e. one >>written in C/C++) is, I imagine, a no-no for most ISPs as would be calling >>C functions from Java (if that is possible). >> >>Has anyone ever implemented such a service? If what I heard about the >>timing issues is actually correct, any comments on how it is possible to >>force Java to behave itself serially/interface Java to something that can >>perform serial comms with an accurate timebase? Please forgive my >ignorance >>on the subject. >> >>Regards, >>Dan >> >>-- >>http://www.piclist.com hint: PICList Posts must start with ONE topic: >>[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads >> >> > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > >-- >http://www.piclist.com hint: The list server can filter out subtopics >(like ads or off topics) for you. See http://www.piclist.com/#topics > > -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

part 1 6764 bytes content-type:text/plain; charset=iso-8859-1 (decoded quoted-printable) Hi, I do agree that it is case-dependent. I have done a project recently with a PIC talking through a hardware TCP/IP stack to a central server and the client based approach solved many security and complexity issues. Incidentally, do you have any idea of volume prices for Tini (say, 10k/annum?) I have not really played around with Java IO very much so am a bit unsure about it's timing capabilities. Speaking to one of our guys who has messed around with the serial side of Java, he said that it wasn't easy to guarantee timed events due to garbage collection and such like. Have you any experience in this area? Regards, Dan (Embedded "Milton Medicinteknik KB, Vikingstad, Sweden" image moved <RemoveMEmilton.medicinteknikTakeThisOuTspamTELIA.COM>TakeThisOuTspamMITVMA.MIT.EDU> to file: 10/10/2001 08:15 pic04474.pcx) Please respond to pic microcontroller discussion list <EraseMEPICLISTspamspamBeGoneMITVMA.MIT.EDU> Sent by: pic microcontroller discussion list <RemoveMEPICLISTKILLspamMITVMA.MIT.EDU> To: PICLISTSTOPspamspam_OUTMITVMA.MIT.EDU cc: Subject: [EE]: Internet devices + java timing Security Level:? Internal Hmmmm.. I agree with you. I am involved with a project at the moment where data is collected with PIC's, and then transferred with a TINI. My id�a was to do approximately what you suggest, but we are many involved and the concept IS TCP/IP in this case. BUT, basically, it is all - of course - very case-to-case dependant. Sven in Sweden. -----Ursprungligt meddelande----- Fr�n: D Lloyd <spamBeGonedan.lloydSTOPspamEraseMEGB.ABB.COM> Till: KILLspamPICLISTspamBeGoneMITVMA.MIT.EDU <EraseMEPICLISTEraseMEMITVMA.MIT.EDU> Datum: den 10 oktober 2001 08:46 �mne: Re: [EE]: Internet devices + java timing >Hi, > >"For the Tini" in what respect? I was thinking of a solution without having >to deploy any TCP/IP devices at the embedded appliance end of the system; >the embedded system would preferably retain just a RS232 capability with >just a plain old modem. This means no extra hardware costs or firmware >changes, too. > >To my mind, it makes far more sense to use some complicated software in a >central place (probably called through or in cooperation with a standard >web server running on a very capable platform) than increasing the hardware {Quote hidden}>end costs of the embedded systems. If the embedded devices have access to a >network (like in a factory), then that is another story. > >Regards, >Dan > > >To: @spam@PICLIST@spam@spam_OUTMITVMA.MIT.EDU >cc: >From: "Milton Medicinteknik KB, Vikingstad, Sweden" ><spamBeGonemilton.medicinteknikKILLspamTELIA.COM> >Subject: [EE]: Internet devices + java timing > > > >To me this sounds like something for the TINI - doesn't it? > >http://www.junun.org/TINI/ > >Sven in Sweden > > >-----Ursprungligt meddelande----- >Fr�n: D Lloyd <.....dan.lloydspam_OUTGB.ABB.COM> >Till: TakeThisOuTPICLIST.....TakeThisOuTMITVMA.MIT.EDU <TakeThisOuTPICLISTKILLspamspamMITVMA.MIT.EDU> >Datum: den 9 oktober 2001 18:43 >�mne: [EE]: Internet devices + java timing > > >>Hi, >> >>Problem: Internet enabling of a device purely having serial comms via the >>internet. >> >>Suppose that a simple device purely supports RS232 and has a modem so it >>may be dialled up. Let's say that such a device operates quite a strict >>timing regime with respect to message turnaround times and such like which {Quote hidden}>>could cause some timing issues if serial comms is not handled to an >>accurate timebase/timeframe. >> >>To Internet enable this same device could, at a simplistic level, require >>the following: >> >>1. Browser running java applet which can talk to a central java >application >>via a socket >>2. Central, muti-threaded java application servicing applet clients. This >>app also perform multi-threaded serial comms to a modem....which is >>attached to the RS232 device. >>3. RS232 device with modem. >> >>This is in the configuration of: >>< Internet > < Server side >>Device side modem link > >>Applet<----TCP/IP----->Server app<--serial-->modem<---------->modem >><---serial---->serial device >> >>In order to contact the serial device, the applet tells the server what to {Quote hidden}>>dial, then the connection is initiated via modem. The server/modem can >then >>hold the telephone line to the RS232 device while data is transferred >>between the system end points. When the user presses the "end session" >>button on the applet, the server drops the modem connection and >>the.....session ends, predictably. Of course, this scenario could be >>handled in CGI/Perl but let's leave it at Java for the moment. >> >>I hear it is tricky to perform timing-sensitive serial comms in Java due >to >>the nature of the beast. To use another app to talk serial (i.e. one >>written in C/C++) is, I imagine, a no-no for most ISPs as would be calling {Quote hidden}>>C functions from Java (if that is possible). >> >>Has anyone ever implemented such a service? If what I heard about the >>timing issues is actually correct, any comments on how it is possible to >>force Java to behave itself serially/interface Java to something that can >>perform serial comms with an accurate timebase? Please forgive my >ignorance >>on the subject. >> >>Regards, >>Dan >> >>-- >>http://www.piclist.com hint: PICList Posts must start with ONE topic: >>[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads >> >> > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > >-- >http://www.piclist.com hint: The list server can filter out subtopics >(like ads or off topics) for you. See http://www.piclist.com/#topics > > -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics part 2 165 bytes content-type:application/octet-stream; (decode) part 3 144 bytes -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Hi, No, I have no pricing for the TINI (other than singles). The garbage handling is done by a set of PIC's in my case. Then - over a RS232-port data is shuffeled over to a RS 232 input on a TINI - and thus to the Net. My part of it more or less ends by the interface of TINI and the application. I don't have any Java (or TINI) experience, this is handeld by others. So - have a modem up my sleeve ;=) Sven in Sweden (further on this maybe you could send privately to RemoveMEmilton.medicinteknikspamBeGonetelia.com ) -----Ursprungligt meddelande----- Från: D Lloyd <spamBeGonedan.lloyd@spam@spam_OUTGB.ABB.COM> Till: TakeThisOuTPICLISTspamMITVMA.MIT.EDU <PICLISTEraseMEMITVMA.MIT.EDU> Datum: den 10 oktober 2001 09:32 Ämne: Re: [EE]: Internet devices + java timing Hi, I do agree that it is case-dependent. I have done a project recently with a PIC talking through a hardware TCP/IP stack to a central server and the client based approach solved many security and complexity issues. Incidentally, do you have any idea of volume prices for Tini (say, 10k/annum?) I have not really played around with Java IO very much so am a bit unsure about it's timing capabilities. Speaking to one of our guys who has messed around with the serial side of Java, he said that it wasn't easy to guarantee timed events due to garbage collection and such like. Have you any experience in this area? Regards, Dan (Embedded "Milton Medicinteknik KB, Vikingstad, Sweden" image moved <RemoveMEmilton.medicinteknikEraseMEspam_OUTTELIA.COM>EraseMEspam_OUTMITVMA.MIT.EDU> to file: 10/10/2001 08:15 pic04474.pcx) Please respond to pic microcontroller discussion list <@spam@PICLISTRemoveMEEraseMEMITVMA.MIT.EDU> Sent by: pic microcontroller discussion list <EraseMEPICLIST@spam@MITVMA.MIT.EDU> To: @spam@PICLISTspam_OUT.....MITVMA.MIT.EDU cc: Subject: [EE]: Internet devices + java timing Security Level:? Internal Hmmmm.. I agree with you. I am involved with a project at the moment where data is collected with PIC's, and then transferred with a TINI. My idéa was to do approximately what you suggest, but we are many involved and the concept IS TCP/IP in this case. BUT, basically, it is all - of course - very case-to-case dependant. Sven in Sweden. -----Ursprungligt meddelande----- Från: D Lloyd <spamBeGonedan.lloydEraseMEGB.ABB.COM> Till: PICLISTspamBeGoneMITVMA.MIT.EDU <RemoveMEPICLIST@spam@spamBeGoneMITVMA.MIT.EDU> Datum: den 10 oktober 2001 08:46 Ämne: Re: [EE]: Internet devices + java timing >Hi, > >"For the Tini" in what respect? I was thinking of a solution without having >to deploy any TCP/IP devices at the embedded appliance end of the system; >the embedded system would preferably retain just a RS232 capability with >just a plain old modem. This means no extra hardware costs or firmware >changes, too. > >To my mind, it makes far more sense to use some complicated software in a >central place (probably called through or in cooperation with a standard >web server running on a very capable platform) than increasing the hardware {Quote hidden}>end costs of the embedded systems. If the embedded devices have access to a >network (like in a factory), then that is another story. > >Regards, >Dan > > >To: .....PICLIST@spam@EraseMEMITVMA.MIT.EDU >cc: >From: "Milton Medicinteknik KB, Vikingstad, Sweden" ><.....milton.medicinteknikRemoveMETELIA.COM> >Subject: [EE]: Internet devices + java timing > > > >To me this sounds like something for the TINI - doesn't it? > >http://www.junun.org/TINI/ > >Sven in Sweden > > >-----Ursprungligt meddelande----- >Från: D Lloyd <.....dan.lloydSTOPspam@spam@GB.ABB.COM> >Till: PICLISTEraseME@spam@MITVMA.MIT.EDU <RemoveMEPICLISTspamBeGoneMITVMA.MIT.EDU> >Datum: den 9 oktober 2001 18:43 >Ämne: [EE]: Internet devices + java timing > > >>Hi, >> >>Problem: Internet enabling of a device purely having serial comms via the >>internet. >> >>Suppose that a simple device purely supports RS232 and has a modem so it >>may be dialled up. Let's say that such a device operates quite a strict >>timing regime with respect to message turnaround times and such like which {Quote hidden}>>could cause some timing issues if serial comms is not handled to an >>accurate timebase/timeframe. >> >>To Internet enable this same device could, at a simplistic level, require >>the following: >> >>1. Browser running java applet which can talk to a central java >application >>via a socket >>2. Central, muti-threaded java application servicing applet clients. This >>app also perform multi-threaded serial comms to a modem....which is >>attached to the RS232 device. >>3. RS232 device with modem. >> >>This is in the configuration of: >>< Internet > < Server side >>Device side modem link > >>Applet<----TCP/IP----->Server app<--serial-->modem<---------->modem >><---serial---->serial device >> >>In order to contact the serial device, the applet tells the server what to {Quote hidden}>>dial, then the connection is initiated via modem. The server/modem can >then >>hold the telephone line to the RS232 device while data is transferred >>between the system end points. When the user presses the "end session" >>button on the applet, the server drops the modem connection and >>the.....session ends, predictably. Of course, this scenario could be >>handled in CGI/Perl but let's leave it at Java for the moment. >> >>I hear it is tricky to perform timing-sensitive serial comms in Java due >to >>the nature of the beast. To use another app to talk serial (i.e. one >>written in C/C++) is, I imagine, a no-no for most ISPs as would be calling {Quote hidden}>>C functions from Java (if that is possible). >> >>Has anyone ever implemented such a service? If what I heard about the >>timing issues is actually correct, any comments on how it is possible to >>force Java to behave itself serially/interface Java to something that can >>perform serial comms with an accurate timebase? Please forgive my >ignorance >>on the subject. >> >>Regards, >>Dan >> >>-- >>http://www.piclist.com hint: PICList Posts must start with ONE topic: >>[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads >> >> > >-- >http://www.piclist.com hint: PICList Posts must start with ONE topic: >[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads > > >-- >http://www.piclist.com hint: The list server can filter out subtopics >(like ads or off topics) for you. See http://www.piclist.com/#topics > > -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics