More Tools . . .

[wiwa64]

Once again i wrote a couple of small DOS programs. This time it's a package of utilities (DOSUTILS) which you might (or might not) find usefull for batch programming (and debugging).

In addition i eventually found a kind soul who was ready to host my software. It can be downloaded from here: http://www.bttr-software.de/products/jhoffmann/. There you will also find the ETHTOOLS, RMENU and NTOOL already known from other threads in this forum.

Even though the layout is not completed yet, you are already welcomed to visit the above mentioned site, download the packages and send me some feed-back, either here in the forum or directly to my e-mail address which you will find in the acompanying documentation.

 

[wiwa64]

Any critics, suggestions, (even applause ) so far? :?:

 

[Ole Juul]

Any critics, suggestions, (even applause ) so far? :?:

The web site looks good.

I'm probably the only DOS user here. Unfortunately, I'm not a programmer so your stuff is of minimal use to me - although a couple of the dosutils might benefit me in some batch file. Anyway, you deserve applause for making a contribution to the DOS world.

 

[wiwa64]

Thank you for your applause. But are you really the only DOS user around? :shocked:

By the way, most of the software doesn't address programmers. Only some of the programs in the DOSUTILS package do. The ETHTOOLS, as well as NTOOL and RMENU are much more related to networking.

 

[TNC]

Sieht gut aus!
English: looks good!

 

[Ole Juul]

Thank you for your applause. But are you really the only DOS user around? :shocked:

I think so. I do know there are several people here with vastly more DOS skills than I, but I don't think they use is regularly. I have noticed that almost all DOS discussions around here end up being MS-Windows discussions. IOW, not pure DOS. It's too bad. Check out Fido:DOS Internet for more focussed discussion. You'll also find people there who have never used anything else and don't even have an internet connection. hehe

By the way, most of the software doesn't address programmers. Only some of the programs in the DOSUTILS package do. The ETHTOOLS, as well as NTOOL and RMENU are much more related to networking.

Well, I found a couple of DOSUTILS which I could use, but who is the ETHTOOLS aimed at? I run a very small network at home (often as little as 5 machines at a time) and use the standard star topology. It's a mixture of routers and switches, just like any other modern network. I can't imagine how I would practically set up a bus topology and assume that only a programmer would do so in order to test something.

Keep the DOS fires burning!
- Ole

 

[mbbrutman]

I think it's all cool and fun to play with. But I know my network topology fairly well, so I'm not going to run analysis tools on it too often. And for the difficult stuff I prefer tcpdump running on my Linux machine.

On an unrelated and slightly off-topic note, I'm thinking about open sourcing my code in the next few weeks. (Or months .. I have some hoops to jump through that normal people would not have to worry about.) The intent is to get more people interested in coding, and thus getting more apps ... Otherwise, we will continue to be the only members of this very exclusive club.

 

[MikeS]

Thank you for your applause. But are you really the only DOS user around? :shocked:

Definitely not; he's flattering himself. I use DOS daily on four different computers and support corporate users who also still use DOS exclusively for certain applications, and I'm sure we're not the only ones.

Will look at your programs when I get a chance, but after 20 years or so I've pretty well got all the tools and utilities that I need and use in my \bin directory; might have another look at Internet access one of these days though. What would be useful for me is a DOS RS-232 <> TCP/IP bridge program like tcpser or BBS Server; anything like that around anywhere?

 

[wiwa64]

. . . but who is the ETHTOOLS aimed at? I run a very small network at home (often as little as 5 machines at a time) and use the standard star topology. It's a mixture of routers and switches, just like any other modern network. I can't imagine how I would practically set up a bus topology . . .

Well, it depends.

First, "bus topology" refers to the logical organisation of the network, which is not necessarily identical to what you can see when looking at the cables. A moderately old 10BaseT ethernet, centered around a hub, for instance, looks like a star topology but it actually behaves like a bus topology. The essential point is, whether one node can "see" all others and listen to all data traffic or whether some instance, like e.g. a switch, isolates it from foreign data traffic.

As the various programs of the ETHTOOLS package can only analyze data they are able to "see", they might be of limited use in a perfectly isoltated environment, while on a bus they can do their job. But this difficulty is not specific to my programs, every network analyzer will face this problem.

So why don't you just try it? If you have a packet driver installed anyway, all you have to do, is start the programs and see what happens. If they don't see any data traffic, which they would tell you, then you are in bad luck, but you didn't loose much as well. If they show results, then it is up to you, whether they give you valuable hints or perhaps you are just curious about what is going on in yout network.

As we all know, DOS is not capable of multitasking. Therefore you cannot run an analizing tool and a network application on the same machine at the same time. But hooking a (cheap) second computer to the same cable, to listen what the others "talk about" is possible. Perhaps it might help inserting a hub between the repective node, the "surveilling computer" and the uplink. Just as a suggestion . . .

By the way . . . while testing the tools, i found a nasty bug in my own small network at home. I accidentilly misconfigured the netmask of one of the nodes in such a tricky way, that it did not completely prevent the respective machine form accessing the net, just the shared drives did not show up correctly on the other mashines. Such an effect doesn't lead you to the suspicion that TCP/IP couldn't be set-up correctly, as ping and all the other standard tools worked fine. Only when inspecting the results of ETHWHAT and trying to understand what was going on, i got the decicive hint, to take a closer look at the netmask.

 

[Ole Juul]

Definitely not; he's flattering himself.

Hehe, well I certainly didn't intend it that way. I use DOS because of my disabilities and although I also use other OSs, I pretty much only use them in the intended way. I am stuck with DOS because it is simple enough that I can use it in the way that *I* intend. It's mostly a matter of lack of skills, or perhaps even intelligence, and that is not actually very flattering if you ask me - but thanks for suggesting it.

I use DOS daily on four different computers and support corporate users who also still use DOS exclusively for certain applications, and I'm sure we're not the only ones.

You are one of the few regular users here. I think there may be a couple of others. The other two Mikes come to mind - and of course wiwa64.

Ole Juul: I do know there are several people here with vastly more DOS skills than I, but I don't think they use is regularly.

I am well aware of the DOS skills people have on this forum, but that does not mean that it is their OS of choice or that they even do their work on it. Note the response that wiwa64 got for this and previous software announcements. Also, Mike Brutman's software is perhaps more useful to a wider range of people (no offense wiwa64 ) and even that got only a couple of people willing to look at it. How many people here use DOS daily? When I say I'm probably the only regular DOS user here I don't think I'm very far off.

 

[Ole Juul]

Well, it depends.

First, "bus topology" refers to the logical organisation of the network, which is not necessarily identical to what you can see when looking at the cables. A moderately old 10BaseT ethernet, centered around a hub, for instance, looks like a star topology but it actually behaves like a bus topology. . .

Thanks for the explanation, that makes sense to me now.

So why don't you just try it? If you have a packet driver installed anyway, all you have to do, is start the programs and see what happens.

Well, of course I already tried it. I try everything that comes my way. I didn't get any response. I do have a lot of flexibility in the setup, so some other time when I rearrange things, I will remember this. Although I haven't had any problems with my networking arrangements, my knowledge is not that deep. This is obviously a good way to do some learning.

As we all know, DOS is not capable of multitasking. Therefore you cannot run an analizing tool and a network application on the same machine at the same time. But hooking a (cheap) second computer to the same cable, to listen what the others "talk about" is possible. Perhaps it might help inserting a hub between the repective node, the "surveilling computer" and the uplink. Just as a suggestion . . .

That is an idea I will keep in mind, as those are both easy ideas to implement.

 

[wiwa64]

. . . I try everything that comes my way. I didn't get any response. . .

Oooops, that looks strange, did you really get no response at all?

Generally the programs are designed in such a way, that they will always give some kind of response. Eather they will inform you, that they are waiting for data to pass by and if no data does so, then this might be the only message you ever recieve, or they will complain about some error, e.g. no packet driver found. The only exception is ETHSEND which by default is silent if everything went well. It only writes a message if told to do so by the /V (verbose) command line switch or if something went wrong.

This is obviously a good way to do some learning.

Curiosity is certainly not the worst motivation to learn something new


How many people are indeed still using DOS? I'm aware, that DOS has lost almost any importance in the commercial context, but there still are DOS forums and some of them still have "data traffic" (even relevant one, not just the usual "watch beautifull teens" messages). And there even is an ongoing project to re-implement a free version of DOS (FreeDOS). But where are all those people?

And what about countries where not so many people can afford the latest hardware, that is required to run Windows? (And even Linux is more demanding with respect to the hardware, than DOS) So where are they all?

 

[MikeS]

Well, for web browsing, email, video, music etc. I use Windows, but for data collection, manipulation & conversion, hardware hacking, working with old computers, games, mirroring clients' DOS systems, etc. I use DOS6.22 or 7.x

Booting and running DOS on a 2+ GHz system is a refreshingly responsive experience compared to the bloated current OSes, not to mention its simplicity and ease of organization and maintenance.

 

[Ole Juul]

Well, for web browsing, email, video, music etc. I use Windows, but for data collection, manipulation & conversion, hardware hacking, working with old computers, games, mirroring clients' DOS systems, etc. I use DOS6.22 or 7.x

I mostly use Linux for browsing, audio, and video. For all writing and most other things it's DOS. I've only been using 6.22 but since I basically only use the kernel and my own utilities, I've thought it might be a good idea to use another DOS kernel. Somehow though, I've never gotten around to trying it because everything works well and perhaps it's better to leave well enough alone.

Booting and running DOS on a 2+ GHz system is a refreshingly responsive experience compared to the bloated current OSes, not to mention its simplicity and ease of organization and maintenance.

I know what you mean about a refreshingly responsive experience! I use a 133MHz and it's plenty fast. I'm not sure that a lot of people have ever experienced that level of speed. And I'm right with you with simplicity and ease of organization. DOS rocks! I've found a bit of love with freeBSD, but it is orders of magnitude more difficult for an amateur like me.

 

[Ole Juul]

Oooops, that looks strange, did you really get no response at all?

Sorry! no I didn't mean it like that. I should have said that I didn't get anything much that looked like it was of use to me. My "testing" (blush) was quite cursory, but actually did get _some_ response.

I just tried it again and also looked at my setup. I currently have a switch plugged into a router which also has a built in switch. I wasn't sure how a second switch would work, but I can see that this is indeed like a bus topology. ETHWHO appears to work, but I'll have to study it. ETHWHAT is slow and on second try was very slow. Interestingly it only finds 3 out of 4 of the addresses working at the moment, whereas ETHWHO was able to detect the one it didn't find. I wonder what that means. WWWATCH works but will also require some rubbing of brain cells. However, after this second try, I think I might find this useful, or at least fun. I only have to type "edc ethtool" to access them on my HDD!

How many people are indeed still using DOS? I'm aware, that DOS has lost almost any importance in the commercial context, but there still are DOS forums and some of them still have "data traffic" (even relevant one, not just the usual "watch beautifull teens" messages). And there even is an ongoing project to re-implement a free version of DOS (FreeDOS). But where are all those people?

I think we have a different and long discussion ahead on this one. I do see DOS in a commercial context. The large Canadian hardware franchise, Home Hardware, uses a DOS front end on their computers. I've seen the same in other businesses. I am also aware of the DOS forums and such. I've been involved there for years. The "DOS Internet" forum is alive as we speak. (Come join us!) I think there are quite a few hobbyists (particularly here) who use DOS for fun. I know there are quite a few of the bulletin board sysops who run dos software, though sometimes in a virtual environment. There also seem to be a number of people who run DOS under Windows. That to me is not DOS because it is removed from the hardware and has external network support. Many people will differ with me on that opinion. Anyway, then you have a few people who run pure DOS as a practical everyday OS. I've noticed that quite a few blind people fall in that category. It is not always clear BTW that a person one is talking to on a BBS is blind - particularly when they have extensive hardware skills too.

And what about countries where not so many people can afford the latest hardware, that is required to run Windows? (And even Linux is more demanding with respect to the hardware, than DOS) So where are they all?

Every time I see discussions about hand-me-down hardware to third world countries I notice one thing - they are not interested. They want modern and they want MS-Windows. Particularly those of us who run DOS and some people in the Linux camp, know how much you can do with old equipment, but that is not what is wanted. They want Youtube and VoIP and stuff like that, because that is what they think a computer is supposed to be. Of course you can do YouTube on a P3 or maybe even older, but that is still a long way from DOS. I think that the DOS we see around is people like us and for industrial use. Not for "desktop" computing. To sum up, I think there are a few hobby users, a few professional users, and very few day to day users.

 

[wiwa64]

ETHWHO appears to work, but I'll have to study it. ETHWHAT is slow and on second try was very slow. Interestingly it only finds 3 out of 4 of the addresses working at the moment, whereas ETHWHO was able to detect the one it didn't find. I wonder what that means.

Theese two programs watch the data traffic and try to evaluate various facts, from what they "see". As a consequence, they cannot deliver results faster than the data actually comes in.

From this perspective, ETHWHO has a somewhat easier job, as it only has to count just about every data package coming by, whereas ETHWHAT needs to see specific data, rather a combination of specific data packages to be able to draw its conclusions.

Why ETHWHAT doesn't see one of the nodes, is indeed an interesting question. Could you be a bit more specific?

 

[Ole Juul]

Why ETHWHAT doesn't see one of the nodes, is indeed an interesting question. Could you be a bit more specific?

My topology is simple. The active nodes right now are:
102 - a Kubuntu machine not being used this minute
101 - another Kubuntu machine running internet radio
50 - a printer
119 - a freeBSD machine idling.
109 - DOS running ethtools.

101, 102, and 50 are plugged into the router. The DOS machine (109) and the BSD (119) are plugged into a switch which is plugged into the router.

Here are the results of two tries:

Total packets: 28       (  0% lost)        Speed:    0 KBit/s

EN_II: 28     IPX: 0      IP: 21     ARP: 1      RARP: 0      Others:  6
SNAP:  0      IPX: 0      IP: 0      ARP: 0      RARP: 0      Others:  0
802.2: 0      IPX: 0                                          Others:  0
802.3: 0      IPX: 0                                          Unknown: 0

Guess:  Netmask=255.255.255.128  Network=192.168.1.0

00142A:34D0FF  192.168.1.102
0050BA:C1DBA2  192.168.1.101
0018FE:825BF3  192.168.1.50
000255:711AD2  192.168.1.119
-----------------------------------------------------------------------
Total packets: 24       (  0% lost)        Speed:    0 KBit/s

EN_II: 100%   IPX:   0%   IP:  75%   ARP:   0%   RARP:   0%   Others:   25%
SNAP:    0%   IPX:   0%   IP:   0%   ARP:   0%   RARP:   0%   Others:    0%
802.2:   0%   IPX:   0%                                       Others:    0%
802.3:   0%   IPX:   0%                                       Unknown:   0%

Guess:  Netmask=255.255.255.128  Network=192.168.1.0

0050BA:C1DBA2  192.168.1.101
0018FE:825BF3  192.168.1.50
00142A:34D0FF  192.168.1.102

The first try showed only the first three nodes as I had seen before, and then I did a random ping from the fourth one (119) just to see if it would wake things up and then ETHWHAT showed that one as well.

I decided to repeat the experiment and ran the ping from 119 right away but got the same result. I waited a minute and ran the ping again - still nothing. After another minute I stopped.

Two more tries and once it showed the 119 after I ran a ping from it. The last time I waited many minutes and ran a ping from 119 several times while waiting yet it didn't show up. So, it appears to be variable. Perhaps I should be doing something more intelligent here.

 

[wiwa64]

Well, the bad news first: I still have no idea, why node 119 is sometimes seen and sometimes not.

The good news is, as i see it, that your example shows (provided you really got a switch and not a hub) that a switch does not (generally) prevent a usefull application of the ETHTOOLS. Apparently data packets from nodes "behind" the switch can still be seen by other nodes, even if they are not directly addressed.

What is not clear yet, is the question whether the switch may perhaps cause only some packets to be seen and others not. Perhaps this is the answer to the ping question (but this is just a possibility, not a definitive truth). Does it make a difference, whether the 119 is the source or the destination af the ping? In other words: Does it make a difference whether you issue the ping request on the 119 trying to ping some other mashine or whether you try to ping the 119 from some other maschine?

Looking at your example, two more questions come to my mind: Why does your router not show up? Did you not have any data traffic with the internet, while you made the tests?

And what are those non-ip frames, counted as "others"? Perhaps they come from your internet radio. By using the cursor keys, you can highlight each of the individual fields in the upper statistics part of ETHWHATs display. Selecting the "others" field will cause the program to diplay more details about this category. Hitting the <HOME> (or <Pos1>) key will return you to TCP/IP analysis.

If this doesn't help to make things clearer, you could repeat the "experiment" recording the data packets using ETHDUMP. Preferably you should write the packets to a file, by specifying the /F command line switch. Later you can examine theese data with ETHSHOW. This way you can inspect all details of the packets and 28 packets are an easily manageable amount of data.

 

[Ole Juul]

Well, the bad news first: I still have no idea, why node 119 is sometimes seen and sometimes not.

I ran ETHWHAT several more times and the first time it didn't find 119, but perhaps I didn't wait long enough. The second time it didn't show up after 15 minutes or more but when I did a ping to the net from 119 then it did show up. The third time I let it run for much more than half an hour and when I came back 119 was still not there. However another outgoing ping made it visible right away.

The good news is, as i see it, that your example shows (provided you really got a switch and not a hub) that a switch does not (generally) prevent a usefull application of the ETHTOOLS. Apparently data packets from nodes "behind" the switch can still be seen by other nodes, even if they are not directly addressed.

This being the VCF, it is of course possible that I would be using a hub, but I don't actually think that they've made those for some years now. The switch is a cheap D-Link DSS-5+ and is labeled as a switch.

What is not clear yet, is the question whether the switch may perhaps cause only some packets to be seen and others not. Perhaps this is the answer to the ping question (but this is just a possibility, not a definitive truth).

I can't answer about the switch treating various packets differently, I wouldn't expect so, but it would certainly be of much interest if it was.

Does it make a difference, whether the 119 is the source or the destination af the ping? In other words: Does it make a difference whether you issue the ping request on the 119 trying to ping some other mashine or whether you try to ping the 119 from some other maschine?

Yes, when I do the ping from the 119 to, say google.ca, then it shows up. When I ping 119 from 101, it doesn't make any difference. As I am writing this, I did several more checks to confirm this.

Looking at your example, two more questions come to my mind: Why does your router not show up? Did you not have any data traffic with the internet, while you made the tests?

No idea about the router. It's a common Linksys WRT54GL with original software and on 192.168.1.1. With Linux it's not obvious what is contacting the net. 101 has roughly 50 windows open now (44 in 5 different browsers), 102 is not mine but is more conservative, 119 is BSD and very basic setup with no browser open right now.

And what are those non-ip frames, counted as "others"? Perhaps they come from your internet radio. By using the cursor keys, you can highlight each of the individual fields in the upper statistics part of ETHWHATs display. Selecting the "others" field will cause the program to diplay more details about this category. Hitting the <HOME> (or <Pos1>) key will return you to TCP/IP analysis.

If this doesn't help to make things clearer, you could repeat the "experiment" recording the data packets using ETHDUMP. Preferably you should write the packets to a file, by specifying the /F command line switch. Later you can examine theese data with ETHSHOW. This way you can inspect all details of the packets and 28 packets are an easily manageable amount of data.

Thanks for all the info (and encouragement), I'll play with it some more.

- Ole

 

[Ole Juul]

I did a bit more playing around and the router does show up. Previously the lights on the router have been blinking from the internet radio connection, (to 101) but that apparently didn't make the node show up. If I ping google.ca from the 101 node which is plugged directly into the router, then the router shows up. It didn't show up when I pinged google.ca from the 119 node which is plugged into the switch nor when I ping 119 from 101. Perhaps because that doesn't actually go through the router, but only through the built in switch in the router. Also, the 119 shows up when I SSH into that box from the 101 - although it didn't do it when I pinged it.

If ETHWHAT shows all traffic and does not make any distinction between SSH and PING, then there is something interesting here. Thanks for getting me going on this. It's not really a priority right now, but I'm on my way. The bottom line is that your programs are indeed useful to me.