Issue with an abbreviation RDT that affects close reasons for questions

Question

First, this is about providing accurate close reasons for questions, that are most likely off-topic anyways. If you don't think it is relevant, you can ignore the rest of the text.

First thing some people think when heaaring this abbreviation is: RDT refers to real data transport. I don't know anything about it, besides that it is a protocol to transport multimedia. And as such, I would assume, that the one thing it does not do is any form of ARQ (i.e., acks, timeouts and retransmissions).

So far there were at least two questions closed, since "they refer to protocol above OSI layer 4 and are off-topic".

Now, this particular close reason is incorrect. The problem is that Kurose and Ross in their book, in their infinite wisdom™ decided to use the abbreviation RDT for reliable data transport, and use this term to refer to what I believe everyone else calls ARQ. Apparently there are RDT 1.0, 2.0, etc which refer to Stop-and-Wait, Alternate Bit Protocol, Go-Back-N, and so on. AFAIK Kurose and Ross's book is still recommended by almost all college networking course. So, people ask questions about ARQ and call them RDT.

Now, I don't quite get the policy about questions about ARQ. Homework is offtopic (so such questions are clearly out). However, design or theory of protocols used to operate a network is ontopic, and, understanding basic ARQs belong to theory of any reliable protocol operating either layer 2 or layer 4. On the other hand, since it is clearly basic building blocks theory, not protocols themeselves, redirecting such questions to e.g., Computer Science CE under tag [computer-networks] is probably a better option.

P.S. And also I would like to point out, that texts of both questions did mention basic ARQ mechanisms. So it absolutely obvious that the questions were about an ARQ mechanism and definitelly not about a multimedia transport protocol. So providing accurate close reason was no problem at all.

Answer 1

Hmm. Questions about L1-L4 protocol theory are legitimate, so ARQ questions are basically on topic, unless convoluted by overly academic or certification-oriented ideas.

Since Kurose and Ross sometimes stray in that direction, like you have indicated, I'd tend to put those questions in that corner. Also, if the OP doesn't provide sufficient context for us to make sense of their question, it's quickly becoming off-topic for being ambiguous and unclear.

However, you're making an excellent point here and we should keep that problem in mind - and ask the OP what the question is about exactly before closing it.

Answer 2

First, this is about providing accurate close reasons for questions, that are most likely off-topic anyways.

There are only a few close reasons:

What do you think should be the close reason for any type of question to which you refer? What you are mentioning falls into the second one that covers a wide variety of off-topic questions, and the reason given the OP for that is listed in the box above the question as:

NE is a site for to ask and provide answers about professionally managed networks in a business environment. Your question falls outside the areas our community decided are on topic. Please visit the help center for more details. If you disagree with this closure, please ask on Network Engineering Meta.

I do not see where that is an inaccurate close reason.

---

Now, I don't quite get the policy about questions about ARQ. Homework is offtopic (so such questions are clearly out). However, design or theory of protocols used to operate a network is ontopic, and, understanding basic ARQs belong to theory of any reliable protocol operating either layer 2 or layer 4. On the other hand, since it is clearly basic building blocks theory, not protocols themeselves, redirecting such questions to e.g., Computer Science CE under tag [computer-networks] is probably a better option.

Remember:

Network Engineering Stack Exchange is for asking questions about professionally managed networks in a business environment.

ARQ and other academic protocols with no standard are off-topic because there is no real-world use in a business network.

Asking theoretical questions about protocols is for real protocols (ethernet, IP, TCP, UDP, etc.) used in a professionally managed network, but there is no standard for academic protocols, so the theory can be whatever an instructor claims, and they are certainly not used in a business network.

If you think a question should be asked on a different SE site, then you should add a comment recommending the SE site to the OP.

Answer 3

I have to agree with @Effie. While such questions are almost universally off-topic, they are so because of their academic nature. Stop-and-Wait, Go-Back-N, and Selective-Repeat are all academic topics. And the questions I've seen here all grew out of academic confusion. To be on-topic, one would have to be asking about something being actively used in real world networks.

TFTP would be an example of stop-and-wait. But it's also "above Layer-4". :-)

TCP, in the 70's, was a stop-and-wait system. It became a GBN with windowing and delayed ACK. But today, most implementations buffer out-of-order traffic unlike the academic defintions of GBN (where out-of-order is discarded.) The buffering mechanism is much more like SR, but SR uses individual ack's.