91Èȱ¬

Some comments:

> 2) These are then sent over a network, often not in the right order or over the same line.

I would lose the "not over the same line" thing altogether, it would be rare for internet routing to change during transmission of a file, and certainly not "often".

The "not in the right order" could be refined (for the graphics dept, as you say), because what is likely to cause this is that a packet doesn't get through, and therefore it has to be resent, which will happen after some of the proceedings packets have already got through sucessfully, so it will be received out of order. (and recombined as in (3)).

Some thoughts.

Packet switching:
1) Packet switching works by taking any information, a picture or email for example, and breaking it up into many small digital pieces, just as a jigsaw can be broken in many separate pieces. These pieces are known as packets.

2) The packets are then sent across the internet by many different routes, but all of them know they have to end up at the same destination.

3) At the destination the packets are put together again, just like jigsaw pieces, so the information becomes a complete whole again.

4) Packet switching is a good way for computers to talk to each other, as it is quicker to send many small pieces of information by several routes than have many large pieces queue to travel along the same route.

Web filtering and blocking:
There is controversy over the terms used for blocking access to sites.
e.g.

It's worth making the point that different countries block access to sites for very different reasons.
China and Cuba come to mind as examples of state censorship that would be seen as unacceptable in western democratic countries.

It's may also be worth giving concrete examples as to why some emails (junk, malware etc) and sites might be seen as being undesirable, and who might have decided to undertake the filtering and blocking.

This can range from parents and schools, individuals (using Firewalls etc), to email services offered by Yahoo or MSN, search services such as Google filtering out 'adult' sites from results etc and so on up to national government level.

Some thoughts on sense and right.

"Packet switching is the perfect tool for computers to talk to each other"

Part of the problem of us trying to understand Internet architecture is that we often anthropomorphize computers. Although conceptually this may be useful, computers do not talk, saying they do infers a human trait. Saying that computers send signals to each other, would remove the anthropomorphic "talk" trait.

Web filtering and blocking:

1) All internet communication works by one computer connecting to another computer, splitting the data into packets and sending them on their way to the intended destination.

A key thing here is the computers never actually "connect", the computer is connected to a gateway and the gateway relays the packets via signals to its gateway and so on. The 2 computers never actually connect. That is a misconception, perhaps easier to visualise, but not accurate. They do not actually talk to a machine called www.bbc.co.uk the process relies on all the devices between the computer and www.bbc.co.uk to be "addressed/configured" correctly. This introduces a lot more conceptual complexity into the equation, but does highlight that machines are oblivious to how the information is retrieved or where it is retrieved from.

I agree with 1squirrel2 on the point relating to ack sequencing. One does not need to know there is an error correction mechanism underlying the process to understand the fundamental concepts of how traffic flows on the Internet. It makes it less simplistic, considering the basic concepts that are trying to be put across here, it seems that introducing error correction in the initial description/graphic may confuse the concept. Introducing it after the basic description/graphic has been presented may be enlightening though.

Regarding web filtering and blocking in general it may just be enough to explain that firewalls and security devices can stop any traffic to or from somewhere, based on any alphanumberic patterns. If the letters bbc.co.uk occurs in the packet seqeunce DO NOT ROUTE, etc. Once again going into who can block what may muddy the concept.

I do not agree with SheffTim on introducing the concept of sending packets via multiple routes, once again this introduces a layer of complexity that is not required to achieve a basic understanding of the concept.

Once the basic concepts are presented, then go into error correction, etc. Indeed I just read that this is programme two - breaking the web. In that case my previous comment is still valid I believe, perhaps more so as not depicting computers as being "connected" to the other shows the large number of points/devices in between which can go wrong or be compromised along a route to somewhere.

I'm not sure if you can get away with it without dealing in the layering of protocols. The layering concept abstracts details to the relevant level making it easy to write programs that talk to other programs. Certainly at the lowest levels packets scoot around the network mixed in with other packets heading to different final destinations and occasionally getting lost.

However at the layer above IP you can treat network connections as a simple pipe between two machines which doesn't loose any information*. This allows web browsers and servers to converse with each other without having to worry about the details and mechanics of how a packet gets somewhere. A typical transaction between network clients is actually fairly easy to follow, look at the out of LiveHTTPHeaders on Firefox if you want to see whats going on under the hood.

The understanding of these multiple layers is certainly useful when you get to deal with the next topic of filtering where different attacks work at different levels of the network.

I understand if you don't want to deal with the next topic of tunnels and working around those attacks for a general audience program :-)

* Well for TCP/IP at least...

I agree with earthgecko, #3, that if your aim is to try and convey concepts to a non-technical audience, some of whom may have never used the Web (30% of the UK population) then you have to keep it as simple as possible.
Many find computing off-putting because it IS technical.

Digital refusenicks:


However re. the discussion above as to whether it is important or not to mention multiple routes - if you don't mention them then why mention packets at all?

I think you're trying to explain the advantages of packets and packet switching; not merely that computers communicate and send signals to each other.
The advantage of packets is that they can be sent off down different routes simultaneously - therefore you have to mention multiple routes.
Otherwise you could just say computers can send information to each other and leave it at that.

However you do it, the technologically literate will groan - but they already understand it.

I thought this a clear, short animated explanation.

I just want to thank you all for your comments on the draft script for programme two explanatory graphics. They've been really useful in avoiding some basic errors on our part and trying to get key details right (such as not using graphics of computers connected to each other without intervening gateways/routers/servers. We are pitching them at a general audience and we want to get across key concepts succinctly and simply but what we say and show has to be accurate.

The simple animation of packet switching that SheffTim recommends is actually the one I pointed our graphics people to as a good example of what we need.
Frank Hanly - director prog 2 Digital revolution

91Èȱ¬ PC Filtering (simplified):

Web Content -> PC -> Check address / content against filter DB on remote server -> PC

The point is, many of these systems don't rely on an intermediary system to filter; they check and withold as they are accessed.