Ed,

More profuse THANKS for all the education you are giving me on this! Again I am printing for later think-through.

My setup is obviously far less than optimized. I am not really bothered by that becuase I only actually use the network facilities infrequently and more just need it to work when I need it.

The cable remians my prime suspect and some corruption caused by MS Combat FltSim install is my second suspect.

Thanks,

Jim N


>-snip-
>
>>That's interesting, to be sure. I know zilch about this stuff and it is the net boys at the office who instructed me to have TCPIP in the Network N'hood. They were emphatic on this, so I listened. The connection I use there is dial-up over ISDN.
>
>If you're routing to another LAN via a WAN connection, then you may need TCP/IP; as mentioned elsewhere, NetBEUI is a non-routable protocol
>
>>It was definitely MS Win98 Help which said to be sure that IPX was there if networking with Win95 systems. I don't know how we poor folks are supposed to get along with 'help' like that!
>>
>
>I haven't seen that note; as mentioned, you can do NetBIOS over IPX/SPX, but unless there's a specific reason, like NetWare (and with more recent versions of NetWare, you can use TCP/IP instead of IPX/SPX), introducing another protocol stack doesn't do much for you, and slows down network packet processing.
>
>>>FWIW, unless you have a need to exchange IP packets between stations on your LAN, IP may not be necessary. It's easy enough to bind TCP/IP to a dial-up adapter; if the packets need not travel to other local stations, there's no need to bind IP to both the dial-up adapter and the LAN. In fact, it would be quite likely that a degree of complexity would be added that might be unpleasant, where your LAN IP address (a static, permanent addres since you don't have a machine to provide DHCP services) and your IP address on your dial-up adapter would be different (it's likely that any dial-up connection dynamically receives an IP address each time you make a connection) adding to the troubleshooting task.
>>>
>>Yea, but I would surely need for the net folks at office to know how to do this, else I'll lose any "support" they can provide.
>>I know nothing about addresses, permanent or otherwise, for the little LAN. I had thought that the NICs had some kind or address in them which did that job.
>>
>
>That's the MAC address, which uniquely identifies the NIC; the IP address is a logical address which identifies your membership in the IP network. It serves as an aid to routing, identifying your network and subnetwork, making it easier to do logical routing. The MAC address is a long unique hex string which doesn't aid in routing; it's an identifier, like a surrogate key. The IP address is usually a 4 octet (fancy name for a number between 0-255) number in dotted form, something like 144.152.96.50 (it's not a real IP address for anything I work on, so don't even bother) that, together with the subnet mask (another 4 octet id that tells how to determine if another IP address is in your local subnet) serve as a basis for routing packets via your gateway.
>
>>>The principle arguments in favor of TCP/IP and IPX/SPX are that the two protocols are routable; they can be carried across multiple LAN segments and routed to their destinations. NetBEUI is a broadcast-based protocol; it is designed to operate over a single segment, and not to be passed across a router, brouter or gateway to a different segment. It is as noted, a bit chatty (the stations under NetBEUI broadcast pckets to make themselves known, making it easy to identify NetBEUI participants with relatively little overhead. With a large number of stations, it generates a significant amount of traffic; you've got three nodes. Packets are small and have relatively little overhead in each packet, since routing and forwarding isn't an issue - it's nice and fast in small LAN environments. It's a simple stack, which imposes relatively little load on each PC as well - with either IP and NBT or IPX/SPX and NetBIOS over IPX, there are multiple software layers involved in processing
>each
>>>packet. There is no need for WINS resolution (the translation of NetBIOS names to IP addresses, the biggest issue for implementing NBT without an NT box to provide WINS services) since all stations use their NetBIOS name as their station identifier in packets.
>>>
>>>I'd recommend the study guides for Microsoft's Network Essentials test - it covers the basics of protocols, media and network troubleshooting, and is a good general base to build on if you ever intend to get an MCSE. Any number of good books on the subject are available and are not heavily theory-oriented; the Microsoft curriculum is quite good.
>>>
>>Clearly I have tons to learn in this area. Without the time to do it! I'll have to look for some literature with the basics in it and hope that is helpful.
>>
>
>Maybe an Evelyn Wood speed reading course would help... :-)
>
>-snip-