07-03-15 09:55 mark lewis wrote to Ed Vance about SCSI Adapter Startup   
      
    ml> @MSGID: <5597083B.14244.windowsa@capcity2.synchro.net>   
    ml> 02 Jul 15 11:05, you wrote to me:   
   Howdy! Again Mark,   
      
    ml>> PATA (old style) and SATA (new style) are both IDE... IDE   
      
    EV> I knew the P meant Parallel and the S meant Serial.   
    EV> I didn't know both xATA HDD's were IDE.  Thanks!   
      
    ml> welcome :)   
      
    EV> I've always thought a Parallel connection was the way to go. I'm   
    EV> thinking about back in the early days Printers were RS-232 Serial, but   
    EV> later most all Printers used a Centronix Parallel connector. Until USB   
    EV> came along.   
      
    ml> yup... parallel requires "latch queues" so that each bit is   
    ml> placed into a latch and held while the others are loaded into   
    ml> their latches... then they are all released at once to travel   
    ml> in parallel... loading the "latch queue" takes time and adds   
    ml> some overhead... "latch queue" isn't what it was really called   
    ml> but it is the closest thing i can think of right now...   
      
    ml> think of parallel as the big toll booth line across the   
    ml> highway... think of the highway leading up to the toll booth as   
    ml> a single lane and the highway leaving the toll booth as a   
    ml> multilane highway with as many lanes as there are booths... as   
    ml> each vehicle approaches, they are directed to the first empty   
    ml> booth on the right so they fill the booths from right to   
    ml> left... they are held in the booth until all booths are full   
    ml> and paid... then they are all released at the same time to   
    ml> travel the next part of the highway in the lane they are   
    ml> currently loaded in... there is no lane changing... when the   
    ml> first group leave the booths, the next ones are queued in the   
    ml> same fashion from right to left to pay and hold until all are   
    ml> loaded and ready... that's how serial is converted to   
    ml> parallel...   
      
   In my thinking about Dot Matrix Printers, I always thought Parallel   
   was faster because of seperate signal lines to each solenoid.   
   On every cycle the proper pins would go BANG! against the Ribbon to   
   make 'their' mark on the paper.   
      
   Serial Dot Matrix Printers would have to wait 8 cycles, plus maybe a   
   START and STOP cycle so the selected pins would move towards the paper.   
      
   Anyway that's my way of thinking about Serial versus Parallel.   
      
      
   ... An unemployed Court Jester is nobody's fool.   
   --- MultiMail/MS-DOS v0.49   
   --- SBBSecho 2.12-Linux   
    * Origin: CCO BBS - capcity2.synchro.net - 502-875-8938 (1:2320/105.1)