On Fri, 19 Aug 2011 16:32:22 +0100, Invisible wrote:

>>> ...but it was somewhat baffling to walk into a classroom and discover
>>> that I know more about the subject than the lecturer does.
> 
>> Oh c'mon now!  Anyone who can think his way out of a wet paper bag has
>> had this experience repeatedly. We're just socially trained not to say
>> so.
> 
> What, that the person being paid to teach you something knows less about
> it than you already do?

Yes, it's not that uncommon.  In my case, it was an assembly language 
class I took in college - it started from how to use MASM, but I'd 
already used TASM quite extensively - but the course was a required 
course in my degree program.

I've also run into that in a number of technical courses I've taken over 
the years - I took a course on Advanced NDS Troubleshooting from Novell 
years ago, but I'd already co-written a book on the topic (in fact, we 
gave a couple copies away in class).  I learned a few things, but mostly 
they were things not in the course.  The instructor was good, but a few 
years ago I was teaching a similar class, and the instructor from that 
first class was a student of mine - and he learned some things he didn't 
know.

Jim

Post a reply to this message

On 19/08/2011 04:41 PM, Darren New wrote:
> On 8/19/2011 8:16, Invisible wrote:
>> OK. So exchange to exchange links and stuff?
>
> That, carrying text messages sometimes, control backplane, DOCSIS, DSL,
> etc etc etc.

> I.e., ATM is used over fiber (usually in turn over SONET) wherever in
> *your* house you are using ethernet.

OK. Is that for legacy reasons, or because ATM is actually good at 
something?

> Honestly, IP is a pretty sucky and limited technology, except for its
> flexibility.

Pffaaahahaha!

"Yeah, IP really sucks, except for being really flexible." Yes, because 
flexibility is a really sucky thing to have.

> It's both inefficient in transmission and extremely
> difficult to manage well.

In what way?

> You couldn't possibly use it internal to the
> phone company for routing or addressing. (For example, one area code
> (aka "city code") in the USA supports more phone numbers than all of
> IPv4.)

Um, excuse me?

The entire population of Kansas is only 2 million people. The IP address 
space is 4 *thousand* million unique addresses. So unless each person 
has a thousand telephone numbers, you don't have a problem.

> But it's everywhere, for the same reason that C is everywhere:
> It's so limited you can layer it on top of pretty much any underlying
> transport.

I still don't see the problem. Then again, IP is the only thing I've 
seen that can handle more than 100 nodes at once...

> For example, one of my colleagues was making fun of ISO's CMIP for being
> connection-oriented, because having trouble getting a connection is one
> of the most common reasons you'd use SNMP. I said "they just dedicate a
> physical connection to it. The smallest bundle going into any switch has
> 900 physical pairs, and usually closer to hundreds of thousands." At
> this point, the student was enlightened.

I'm not though.

Post a reply to this message

On 8/19/2011 8:41, Darren New wrote:
> (For example, one area code (aka "city code") in
> the USA supports more phone numbers than all of IPv4.)

Actually, I did that math wrong. Nevermind. :-)  But certainly just USA 
dialing supports more numbers than IPv4, let alone the six orders of 
magnitude added by international dialing.

-- 
Darren New, San Diego CA, USA (PST)
   How come I never get only one kudo?

Post a reply to this message

Mike the Elder <nomail@nomail> wrote:
> Invisible <voi### [at] devnull> wrote:
> > As part of the drive to empty our house of useless crap, I was going
> > through some of my old college work last night. Ah, the memories.
> >
> > ...but it was somewhat baffling to walk into a classroom
> > and discover that I know more about the subject than the lecturer does.
> Oh c'mon now!  Anyone who can think his way out of a wet paper bag has had this
> experience repeatedly. We're just socially trained not to say so.

  When I went to the university, I don't remember any obviously incompetent
professors. Some teaching assistants were, but that's completely different
(because they were often just students like everybody else, usually just
a few years senior).

  Most (if not all) of the professors related to computing science that
ever taught me were much more competent and knowledgeable about programming
than I am today. Perhaps the only field where there was slightly less
abundant expertise overall was computer graphics.

-- 
                                                          - Warp

Post a reply to this message

Am 19.08.2011 17:16, schrieb Invisible:

>>> How about ATM, is that still used?
>>
>> It's used quite a lot internally to the phone company. You're probably
>> using it right now.
>
> OK. So exchange to exchange links and stuff?

If you still have a proper "analog" or ISDN phone, chances are you 
actually have voice-over-ATM except on the "damned last mile".

Post a reply to this message

On 8/19/2011 8:49, Invisible wrote:
> OK. Is that for legacy reasons, or because ATM is actually good at something?

ATM is a lot less legacy than IP is. Yes, of course it's actually good at 
something.

> "Yeah, IP really sucks, except for being really flexible." Yes, because
> flexibility is a really sucky thing to have.

I didn't say that. I said it sacrifices other things in preference to being 
flexible. There's no resource allocation. There's no way to force a 
particular route thru the network (well, there is, but nobody actually 
implemented it in their switches), the addressing sucks for large networks, 
the address space (in IPv4 at least) is exceedingly limited, remote 
management of hardware is extremely limited, there's no access control, 
admission control, or decent rate regulation other than actually dropping 
packets, etc etc etc.

Heck, it doesn't even do roaming, which cell phones have been managing for 
10 years.

>> It's both inefficient in transmission and extremely
>> difficult to manage well.
>
> In what way?

See above. Every packet has the full source and destination address, and 
there's no information anywhere about the physical network.

> The entire population of Kansas is only 2 million people. The IP address
> space is 4 *thousand* million unique addresses. So unless each person has a
> thousand telephone numbers, you don't have a problem.

Yeah, I realized I mentally screwed that up, and corrected it in the next 
message. :-)

And remember, tho, that IP needs an address for every interface, not for 
every endpoint. So if I make a phone call from my phone to the switch to 
your switch to your phone, that would take 4 IP addresses using IP.

>> But it's everywhere, for the same reason that C is everywhere:
>> It's so limited you can layer it on top of pretty much any underlying
>> transport.
>
> I still don't see the problem. Then again, IP is the only thing I've seen
> that can handle more than 100 nodes at once...

You... are unfamiliar with phone lines?  A *small* phone switch handles a 
half a million end users and hundreds of thousands of trunk lines.

>> For example, one of my colleagues was making fun of ISO's CMIP for being
>> connection-oriented, because having trouble getting a connection is one
>> of the most common reasons you'd use SNMP. I said "they just dedicate a
>> physical connection to it. The smallest bundle going into any switch has
>> 900 physical pairs, and usually closer to hundreds of thousands." At
>> this point, the student was enlightened.
>
> I'm not though.

Let's just say that managing and programming hardware is rather different 
depending on whether you're dealing with a lossy packet-switched network or 
a connection-oriented network to which you can dedicate actual physical 
connections for programming and management. Mocking configuration of the 
latter because it would work poorly over the former is missing the point.

-- 
Darren New, San Diego CA, USA (PST)
   How come I never get only one kudo?

Post a reply to this message

On 8/19/2011 5:00 AM, Invisible wrote:
> As part of the drive to empty our house of useless crap, I was going
> through some of my old college work last night. Ah, the memories.
>
> I still chuckle at my first few programming assignments. We were
> learning to use Borland Turbo Pascal 5.5 for MS-DOS. The guy teaching us
> only found out two weeks earlier that we was taking the class, so he had
> two weeks to learn Pascal out of a book. Frankly, they should probably
> have just given us that book! (Also, his name was Frank.) He wasn't
> /bad/ at teaching, but it was somewhat baffling to walk into a classroom
> and discover that I know more about the subject than the lecturer does.
>
> The comments on my first assignment indicated that it works well, "and
> you've even added a few extra procedures you weren't asked for" (in
> other words, I factored the problem more thoroughly), "but comments are
> rather sparse". Considering that the entire source code print out was
> literally 3 pages long, there's not a hell of a lot to comment on. No
> non-obvious design decisions or tricky invariants to document. If you
> understand Pascal, the code is pretty much self-explanatory. It's a
> first-year college assignment FFS!
>
> My second assignment is even more amusing. The actual phrase the
> lecturer used was "...the program works, but is a little overly complex
> for mere mortals like me to understand." Oh, I'm sorry, I thought you
> were supposed to be the frigging *expert* that we've all paid good money
> to be taught by! :-P
>
> In fairness, as well as writing the code for the actual assignment, I
> did also write an entire library module that's triple the size of the
> main program, and which allows me to print nicely formatted output. (I
> forget the exact details now. I think it was just number alignment so
> all the decimal places line up.) But I put it in a separate source file
> and made it really easy to read the main program without having any idea
> how the library actually works. And I put comments in the library
> explaining what each procedure does (not that the names didn't make it
> self-evident anyway).
>
> Then again, splitting a program into more than one module was beyond the
> scope of our course. (!) We were never taught the syntax for doing this.
> (It's not like C where you just write more files. In Pascal the main
> program source file has a different structure to library modules. You
> have to declare the public interface, for one thing...)
>
> This experience of knowing what the lecturer is going to teach us before
> he even opens his mouth turned out to be a recurring theme of my path
> through college and university. At the time I just felt smug for being
> better at Pascal than Frank was. By the time I was in my 3rd year at
> university, I was beginning to feel frankly a little outraged that I was
> being charged vast sums of money (tens of thousands of pounds, more
> money than I will ever own) to listen to lecturers who can't speak
> English properly and/or don't have a damned clue WTF they're talking about.
>
> (Not /all/ the lecturers were this bad, of course. But some of them
> were...)
>
> In one of Frank's other classes, we learned all about binary and
> floating point. The former was of course no longer news to me. Spend 10
> years poking bytes into the control registers of the 6581 (that's the
> MOS Technology Sound Interface Device [S.I.D.] to you) and you learn a
> thing or two about binary, computers, processors, addressing modes, and
> so on. Floating point was all new to me though.
>
> Then there were the lessons on logic gates. (You had to know what they
> are. You weren't required to know what to do with them or why they
> exist.) Again, old news to me, having already spent years playing with
> physical 7400 chips. Taking a harddrive apart was fun though. (And
> probably ****ing expensive, I should think. This is the era when "4 gig
> drives don't grow on trees".)
>
> And then there were the Quants lessons. ("Quantitative Methods") This,
> as best as I can tell from my lecture notes, was simply an orgy of
> mathematics. It's all in the bin now, but I skimmed through pages and
> pages of statistics, polynomials, linear algebra, fractal geometry, 3D
> graphics, Polish reversed lists, chaos theory, cryptography, sound
> synthesis, and God-only knows what else. I should note that only about
> half of this stuff is directly related to anything in the syllabus. The
> other half is just me doodling while bored in class. (Something that
> DJK, our lecturer, seemed to actively encourage.)
>
> I actually found the sheets of paper where I first performed the
> derivation of the binomial theorem from first principles. You cannot
> simply /tell/ me that (A + B)^2 = A^2 + 2AB + B^2. I have to know /why/
> this is true. And why is it that (A + B)^3 isn't A^3 + 3AB + B^3? There
> must be a pattern. And damnit, I'm going to find it.
>
> This apparently involved drawing a grid:
>
> +------+---+
> | | |
> | A^2 |AB | A
> | | |
> +------+---+
> | AB |B^2| B
> +------+---+
> A B
>
> I think you can see where this leads. Next I drew the same thing, but in
> 3 dimensions. That took me a bit longer. After that, I wanted 4
> dimensions, but was forced to abandon geometry and simply shift algebra
> by hand. The final expression may be small, but some of the intermediate
> terms are quite large, and it's easy to make mistakes as you repeatedly
> copy and manipulate the terms step by step.
>
> After laboriously tabulating the expansions up to (A + B)^9, it became
> quite obvious what the general pattern was. But what the hell are the
> coefficients? Where do they come from? It took at least another hour of
> shifting algebra around and scrutinising my work before I discovered
> that collecting like terms caused the coefficients to be computed in a
> way exactly matching the definition of Pascal's triangle.
>
> DKJ offhandedly told me that I had discovered "the binomial theorem".
> Consulting the library, I learned that this little puzzle had been
> solved over a century ago. AND FOR NEGATIVE EXPONENTS! >_<
>
> This again is a typical experience in mathematics. Any problem which you
> can imagine has either been solved several hundred years ago, or else is
> impossible to solve. There are no easy unsolved problems. I had the
> feeling of being in the centre of a field of ideas, looking at all the
> familiar, well-known stuff. All the new, unknown stuff would be at the
> far reaches of the field, involving problems so complicated I wouldn't
> even understand the language.
>
> The first time we were in the computer lab, DKJ got us to graph some
> trivial polynomials which supposedly represented various things. (E.g.,
> rather than give has a data set representing the company's monthly
> income, he would think up a cubic off the top of his head and get us to
> tabulate and graph that - an easy task for Excel.) I of course was
> bored, and so I ended up graphing my standard incantations. The harmonic
> expansion of various waveforms. The chaotic behaviour of the logistic
> and lambda maps. And so forth. As I recall, it caused quite a stir among
> the chavs I shared a classroom with. For five minutes.
>
> I recall on another day, we had one formula which yields the company's
> income for each month, and another that yields the expenses for each
> month. We were using this to understand the concepts of profit, loss,
> break-even, etc. This was a classroom exercise, so everybody sat
> laboriously tabulating these formulae by hand.
>
> Obviously, I immediately realised that I could directly tabulate the
> company's monthly profit simply by subtracting one formula from the
> other, simplifying the algebra, and then tabulating that. But then DKJ
> wanted us to compute the profit growth rate. I duly tabulated the first
> few months of profit and computed the difference between consecutive
> rows. But I couldn't help feeling that there ought to be a /pattern/ to
> the results.
>
> After about 15 minutes of experimentation, I managed to successfully fit
> a quadratic curve to my data. Literally, I had a formula that produced
> the exact same numbers as the curve I was laboriously tabulating. And
> it's the damnedest thing: the coefficients of this new formula seemed to
> be /related/ to the original formula. Well, it seemed obvious to me that
> such a relationship /should/ exist, but what exactly was its nature?
>
> While everybody else continued the tabulation of growth rate that I had
> long since finished by halving my workload, I picked formulae at random,
> tabulated their growth, and tried to fit curves to them. This is quite a
> lot of work, and I never did figure out exactly what the relationship
> was. At this point, DKJ came over and informed me that I had just
> invented differential calculus, and showed me the general formula for
> the derivative of any polynomial.
>
> I remember being quite suspicious of the perfect constants in the
> expression. "Is that really 3? Or is it 2.986 or something?" No, it's
> *exactly* 3. Needless to say, I went and found a book which explained
> all this in far more detail than the 5 minute conversation I had in
> class, and now I understand /why/ all of these beautiful patterns are so.
>
> While sorting through all this paperwork, there was one assignment I
> felt compelled to keep. The assignment, quite simply, was to write a
> program to "do graphics". Literally, any sort of program you write which
> produces pretty graphics is OK. While DKJ was explaining on the board in
> excruciating detail how to draw a simple rotating square, I was typing
> away at my computer, which I had already got drawing a rotating
> wireframe 3D cube.
>
> It was around this time I was playing with complex numbers. For years
> I'd been baffled by the descriptions telling me that the Mandelbrot
> formula is z := z^2 + c, but the formula is /also/ x := x^2 - y^2 + a, y
> := 2xy + b. Clearly z = x + yi and c = a + bi. But how do you get from
> one formula to the other? It made no sense, and nothing anywhere
> explained it.
>
> One fateful day, I tried applying the binomial theorem to (x + yi)^2 +
> (a + bi). The result is obviously x^2 - 2xyi + y^2i^2 + a + bi. If you
> replace i^2 with -1 (for that is its defined value), you get x^2 - 2xyi
> - y^2 + a + bi. And if you gather all the terms containing i in one
> group and all the others in another, you get x^2 - y^2 + a and 2xyi +
> bi. Divide the latter through i and its 2xy + b.
>
> Suddenly, it clicked. (No, not my spine!) I spent the rest of the day
> deriving complex arithmetic from first principles. Perhaps the single
> most significant moment was in class the next day. I wrote out the
> Taylor series for exp(yi). Replacing i^2 with -1, I discovered that all
> the terms now alternate between negative and positive. And between real
> and imaginary. And collecting all the reals and all the imaginaries, I
> was shocked to find myself looking at the Taylor series for the sine and
> cosine functions.
>
> At which point DKJ informed be that Euler figured that out almost 300
> years ago. Damnit! >_<
>
> Following our lecture on Polish reversed lists, I wrote a small Pascal
> interpreter. It takes a String containing a Polish reversed list, and
> executes it. You can only include purely real constants and the variable
> names "A", "B", "C", "D" and "Z". The values of these variables are
> passed as arguments to the interpreter function. The interpreter
> executes the expression, and returns the result. It even correctly
> handles complex exponents. (Although they would have to be variables.)
>
> Using this, I added to my graphics assignment a fairly large collection
> of fractal types. I don't think I ever got as far as implementing
> Dijkstra's shunting algorithm (to turn normal algebra into Polish
> reversed lists). But I implemented the usual Mandelbrot and Julia
> fractals, with multiple colouring options (escape time, Z-magnitude,
> epsilon cross, binary decomposition, etc.) And also several other
> fractal types; cubic Mandelbrot, lambda, logistic, lambda-exp and
> lambda-sin.
>
> I read in a book about Floyd-Steinberg error distribution dithering. Now
> remember, we're talking about the era where one of my assignments
> actually asked me about the differences between MGA, CGA and EGA. (For
> you youngsters: That's what we had *before* VGA.) We all did our
> graphics work in VGA mode. So, 16 hard-coded colours available. So I
> went ahead and wrote an FS dithering algorithm, and then adapted my
> Mandelbrot renderer to generate images in 24-bit colour and then dither
> them into 4-bit colour. The results at 640x480 were... well, grainy.
> (Plus I think I had a bug somewhere. The algorithm never worked
> /exactly/ right.) But it was still kinda neat.
>
> I actually /sold/ the code for the Mandelbrot fractal to Graham. Not for
> very much money. And, actually, I didn't give him any code. I just sat
> next to him and told him what steps had to happen, and he coded it. I
> can still remember the conversation. "Now run that loop for every
> pixel." "Wow, for EVERY PIXEL?" "Yep. Now you know why it's so slow!"
> (Mine of course did fancy tracing operations to speed it up...)
>
> My rotating cube went on to acquire hidden line removal. *ahem* OK,
> back-face cull. Which took some considerable research and discussions
> with DKJ to get right. Finally I discovered that it wasn't working
> because it /actually matters/ what order you list the coordinates in.
> O_O Once I fixed that, it worked much better. I then went on to add
> surface illumination calculations, using a simple point source and Phong
> lighting. (Each surface is one flat colour.) With only 4 shades of grey
> to play with, it didn't look so hot. But it was quite neat animated.
>
> I vaguely recall that I did implement something with /real/ hidden line
> removal, but I don't know if it made it into my assignment. I'll have to
> read the sources.
>

> animated spirals and so forth. Needless to say, I got a Distinction for
> that. (The highest mark possible.)
>
> It still makes me chuckle that I enrolled for a computing diploma and I
> spent two terms learning how double-entry accounting works. I've got
> Sage printouts and everything. (That's Sage for MS-DOS, mind you.) I
> still remember the login password. It was "letmein".
>
> Then there's heaps of stuff about the relative merits of Ethernet verses
> Token Ring verses something else. As far as I can tell, that's a format
> war that has long since been won my Ethernet. Does *anybody* still use
> token ring? How about ATM, is that still used?
>
> (Come to think of it, *all* of the networking stuff at college was done
> by Novel Netware. Now there's a name I haven't heard in a long time...)
>
> There was one assignment called "computers in society". I didn't
> actually read the assignment, but the lecturer's comments were amusing.
> "A wonderful, almost poetic report. I would be delighted to receive
> writing of this calibre in my professional life." He then goes on to
> offhandedly mention that I didn't actually cover all of the required
> aspects. And yet, I got a Distinction? LOL!
>
> I briefly flicked through the report I wrote. Obviously produced with MS
> Word 6. One sentence has been highlighted and the guy wrote next to it
> "beautiful!" The sentence reads:
>
> "The computer is to information as the power loom is to fabric."
>
> Aside from that, the report isn't well written at all. The tone is very,
> very informal. Like, if you imagine being given a topic, and just
> monologuing about it to your friend off the top of your head,
> videotaping that, and then rearranging the sentences into a more
> coherent order with a word processor... that's what I wrote.
> Distinction. Heh, well I guess it *is* only college...
>
> Also: If you think my spelling is bad *now*, you have literally no idea
> how abysmal it used to be! (E.g., I can spell "abysmal" now. :-P ) And
> that was when I was a teenager, obviously.
>
> Some of the lecturer's comments weren't so friendly. (Actually, some of
> them weren't legible at all. That has to be some kind of fail, right
> there!) Deanna wrote something about "Task A shows technical brilliance,
> however task B is still not acceptable, even on this second attempt, and
> even though I told you which research sources to use. This just isn't
> good enough Andrew. You simply must improve in this area."
>
> Oh? Oh really? You think so? Do ya? Yeah? Well let's see how you like my
> SHREDDER!!! Muhuhuhuh!!! Take THAT, Selby! Yeah, you like that?! Not so
> smug *now*, are we? Now that your acidic little smartypants voice has
> been minced into a thousand tiny pieces. EAT METAL!!! >:-]
>
> Ahhh. To quote H. Granger, "that felt good".
>
> In summary, all the assignments to do with programming computers or
> solving equations gave me exemplary marks. Everything else brought me
> dismal failure. A pattern that was to be repeated at university, come to
> think of it. Well I guess that's tomorrow's tidying task.
>
> I'm particularly proud of the grade I got for my second programming
> module at uni. I got a D for that. (On an A - F scale.)
>
> You see, due to the faculty secretly moving the notice board and not
> bothering to tell me about it, I missed all of my exams that term. Every
> single one of them. And that meant I failed all my modules and had to
> retake them.
>
> Well, all except one. You see, for my programming module, my coursework
> grades were *so* damned awesome that I actually PASSED THE MODULE
> OUTRIGHT WITHOUT EVEN TAKING THE EXAM. And not only did I pass, I got a
> D. Not an F. Not an E. But a D. Without even sitting the exam. I'm *that
> good*, bitch! ;-)
>
> And then there was that database class. The one where Ian says "There
> are various levels of normalisation. I mean, you've got 2NF, 3NF, 4NF
> and even higher ones. But generally, you only really need to worry about
> Boyce-Codd normal form, which says, ANDREW..."
>
> I stand up from my seat in the front row and enunciate "...every
> determinate must be a candidate key..."
>
> "...thank you Andrew. Now as you can see[...]"
>
> That was a feel-good moment. When even the lecturers know you're a
> damned walking encyclopaedia. ;-)
>
> ...and now I'm feeling all nostalgic about the long-lost days when I
> could just sit in a corner reading about Gaussian elimination or
> something while the lecturer blathers on about the role of middle
> management or whatever. I could do crazy math stuff, and people would be
> *impressed* and stuff. My college notes are literally a fermenting sea
> of equations. Serious work mixed up with doodles and advanced math. I
> have nobody to impress anymore. :'{
>
> (Although... I wonder how much of the praise I received was actually
> justified. Especially at college, when I was in a class full of chavs
> who went out and got drunk *every day*. They'd be hung over for the
> morning classes, and mildly drunk for the afternoon ones. Is it any
> wonder the lecturers though I was brilliant?)
>
>
>
> Go on, admit it. You thought this was going to be about SRAM vs DRAM,
> didn't you? :-P
>
You do realize that about half the people here, including me, read this 
and thought, "Man I really hate this guy for being able to do that 
shit!", right? lol No wonder I can't even figure out some "basic" stuff 
I need for some 3D math. The most I ever derived was some non-standard 
way of multiplying two, two digit, numbers, so it wasn't necessary to do 
all the complicatedly silly stuff the teacher insisted we screw with. :p

Post a reply to this message

On 19/08/2011 05:03 PM, Warp wrote:

>    When I went to the university, I don't remember any obviously incompetent
> professors. Some teaching assistants were, but that's completely different
> (because they were often just students like everybody else, usually just
> a few years senior).
>
>    Most (if not all) of the professors related to computing science that
> ever taught me were much more competent and knowledgeable about programming
> than I am today. Perhaps the only field where there was slightly less
> abundant expertise overall was computer graphics.

Apparently I went to a bad university.

Ironically, our Computer Graphics lecturer actually *was* an expert. I'm 
not sure getting a lecture hall full of people to do matrix 
multiplication on paper is especially productive, but the guy definitely 
knew his stuff.

I can still remember everybody laughing when he said that our assignment 
was to write a program to draw a straight line. But *you* try doing that 
with only integer arithmetic. It's surprisingly nontrivial.

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

>> OK. Is that for legacy reasons, or because ATM is actually good at
>> something?
>
> ATM is a lot less legacy than IP is. Yes, of course it's actually good
> at something.

Well, you know, RS232 is a pretty sucky design. But it's still here. 
That *isn't* because it's good at anything. It's because it's widely 
implemented, i.e. legacy.

>> "Yeah, IP really sucks, except for being really flexible." Yes, because
>> flexibility is a really sucky thing to have.
>
> I didn't say that. I said it sacrifices other things in preference to
> being flexible. There's no resource allocation. There's no way to force
> a particular route thru the network (well, there is, but nobody actually
> implemented it in their switches), the addressing sucks for large
> networks, the address space (in IPv4 at least) is exceedingly limited,
> remote management of hardware is extremely limited, there's no access
> control, admission control, or decent rate regulation other than
> actually dropping packets, etc etc etc.

Most of this sounds like "IP is sucky for managing network hardware". 
Well, no, that's not what IP is for. It's deliberately hardware-neutral. 
The idea is that you have some infrastructure for controlling your 
hardware, and then run IP on the top. IP is for moving data from A to B.

If you asked me what was "sucky" about IP, the one I'd probably pick is 
that the entire design philosophy fundamentally assumes that everybody 
will follow the rules. That means it's trivially easy to read arbitrary 
traffic, and to screw up the network. For example, the way I heard it, 
nobody implements source routing because it makes it far too easy to 
impersonate other people, or do DDoS attacks, or whatever.

> Heck, it doesn't even do roaming, which cell phones have been managing
> for 10 years.

Cell phones do that by being controlled by a single central authority.

> See above. Every packet has the full source and destination address, and
> there's no information anywhere about the physical network.

Ooo, 64 bits, bit deal.

It sounds like you're basically complaining that IP isn't 
connection-oriented. You know, if what you're trying to do isn't 
connection-oriented, that's an advantage, right?

> And remember, tho, that IP needs an address for every interface, not for
> every endpoint.

Well, that's true though, and a bit of a pitty. Still, maybe in 20 
years' time, IPv6 will actually catch on...

>> I still don't see the problem. Then again, IP is the only thing I've seen
>> that can handle more than 100 nodes at once...
>
> You... are unfamiliar with phone lines? A *small* phone switch handles a
> half a million end users and hundreds of thousands of trunk lines.

I'm unfamiliar with the data protocols that enable this to be possible. 
(I was under the impression that everybody runs their voice data over IP 
now anyway, so they only need to maintain one big IP network for their 
voice services *and* their broadband offerings...)

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Orchid XP v8 <voi### [at] devnull> wrote:
> I can still remember everybody laughing when he said that our assignment 
> was to write a program to draw a straight line. But *you* try doing that 
> with only integer arithmetic. It's surprisingly nontrivial.

  Drawing a line with integer arithmetic is quite easy. Drawing it
accurately is slightly more involved. Drawing it *fast* is non-trivial.

  And that's assuming the line has a width of 1 pixel and the drawing is
done without antialiasing.

-- 
                                                          - Warp

Post a reply to this message