Free Metalworking Plans

According to Leon Fisk :

[ ... ]

Hmm ... when I used wget on it, it pulled the whole file in a single try -- and xpdf opens it and reads to the end.

And in xpdf 3.00 (on a Sun Solaris system, fwiw).

But quite useful.

I find myself wondering which magazines these old articles are from. The shaper article showed the year (1955, IIRC), but no hint of the magazine itself.

Enjoy, DoN.

I believe they are from Popular Mechanics, from the days when it was a decent magazine. My father has a box of them from when he was a young man - I loved reading them when I was growing up. Stuff you wouldn't see today, like explosive forming sheetmetal with a homemade gizmoe and a .38cal blank..

When I was young there were four magazines that had articles like the ones posted. The other were Popular Science, Science and Mechanics, and Mechanix Illustrated. I noted that one of the articles posted was from Model Engineer and another was from Mechanix Illustrated.

Dan

Those were the days of lathes and mills within and they had special issues on making machines and projects. I have two of said specials that Dad had and I scarfed when I last visited.

Martin Martin H. Eastburn @ home at Lions' Lair with our computer lionslair at consolidated dot net TSRA, Life; NRA LOH & Endowment Member, Golden Eagle, Patriot"s Medal. NRA Second Amendment Task Force Charter Founder IHMSA and NRA Metallic Silhouette maker & member.

Geoff wrote:

some of them are from mechanix illustrated, however not sure what magazine they are really from but this link should keep you happy for hours

It's not a flakey server, as evidenced by the fact that wget has no problem. For some reason, IE just doesn't know how to download pdfs. I see it all the time on my server. IE needs several tries while all the other clients manage to get pdfs the first try.

Another tip on using wget is to ask it to use a different user agent than the default.

alias Wget='wget --header '''User-agent: Mozilla''''

e.g.

wget --header "User-Agent: Mozilla"

i

NO, it was a server problem. The connection is/was being closed prematurely. Here is the log file Wget generated when I retrieved the file (ignore the quoted material mark, old trick to keep the lines from wrapping):

I'm sure it varied from time-to-time and with the speed of peoples internet connections as to when the server decided to up and close prematurely...

Another neat thing you can do with Wget is find out what the date/time stamp is on a file. Try using this command line (ignore the ">" sign):

Substitute a good url first though. The result will be written to the file "wget-spec.log" in whatever directory you launched Wget from. Works nice to find out if a file has been updated or not before downloading it and a few other things about the server.

I'm glad this helped you out :)

Your log doesn't show that it was a server problem. It only shows that you had some undetermined disconnect problem. No reason given. Might have been the server, but IME, I'd bet something else... like network bottleneck between you and the file you were trying to retrieve. As I said before, my server logs very OFTEN show Windows trying to get a pdf, stopping prematurely, and then immediately trying again. Or sometimes, getting the whole thing, then retrying and only getting a partial file the second and third tries. Anyway...

If wget output proves anything, then here's mine from my FreeBSD machine:

connected.

And here's the one from my Linux (rhymes with Linus) machine:

There you go! No server problem. In both cases, I got it in one go, 5 seconds each.

I just reviewed my apache logs for the last 60 days. ($ zgrep pdf *gz|grep 206) and much to my surprise, all the 206's didn't go to IE, though every single one of them did go to Windows machines. Surprise, but no surprise.

I see in my twoloonscoffee server log one completely broken Windows machine that attempted to get a pdf file

220 times in 6 minutes, getting a few more bytes each attempt... and finally getting the whole thing on try number 221. (I wonder if someone was sitting there left-clicking over and over... in which case, I suppose we couldn't really blame Windows.) I can say with certainly though... Every single 206 error (partial content) on a pdf file was given to a Windows machine.

Agreed. And I've been looking at web logs for a _very_ long time.

Might be worth tweaking the keepalive and or session timeouts, but that can have problems if you have a high volume site.

Since when?

Both of which show that timing and timeouts are likely culprits.

Not unusual at all. The logs I've seen over the last mumble years are consistant with your observations. By the way, there's a tool called "analog" which does a fantastic job of showing this sort of thing in a graphical way. If you haven't seen it, it's worth checking out. Might be in your ports tree already.

Dave

Since at least back as far as Red Hat 4.1. Used to be every distribution had an *.au file (later changed to *.wav) of Linus saying, "Hello, my name is Linus Torvalds and I say Linux as Linux." ... where he pronounces Linux to rhyme with Linus. I check the Debian box, and find no such file in Debian 4.0 though.

Google turns up this penguin video overlaid on that same audio track here though:

The first Linux book I ever bought pointed out that in English Linus is pronounced with a long i, Lye-nus, and most everyone who was a Linux user in the mid 90's said Linux that way... except the ones who were also Windows users. That crowd figured that Win and Lin should rhyme... and the majority started saying it the Windows (wrong ;-) way. From a phonics point of view, how would you justify pronouncing the "i" in Linus and Linux differently?

In early TV interviews, Linus, speaking English, also pronounced it with a long i. Eventually the Windroids got even to him, and last I saw him on TV, he was pronouncing it both ways even within the same paragraph.

Analog shows 268 "206 errors" issued in the monthly report, (220 of which were from a single user) but not for which files or even which file types, or how that correlates to browser or OS or user. Using logs and grep and wc, and even sort when needed, gives much more information than analog.

It's one of those things I've had installed for like, ever... but almost never use.

Um, OK, fair enough, how does he pronounce Linus? Got a link for the .wav? I tried googling but can't come up with it...

I looked through an ubuntu box with the same result.

Ah. Your google-fu is strong.

Right. And then there seemed to be this wave of "those in the know" correcting those of us saying it that way.

Interesting. I'm in the midwestern USA, maybe it's a regional/country thing?

I'm not sure that applying logic to linguistics is, well, logical. SO many counterexamples, after all.

So language is evolving. It has a habit of doing that.

the analog.cf (analog.conf maybe?) file is highly configurable - something like "add all" will give you every possible slice & dice. Depends on if you want to do it once, or regularly. Once, sure, grep this, grep -v that, sort by whatever the other|uniq, sure.

People who own websites like it, management and all that get off on the pretty graphs and so on.

According to Leon Fisk :

[ ... lots of logfile snipped ... ]

Hmm ... here is my re-download (I already had it, but did not save the information from wget):

====================================================================== Fuego:dnichols 16:57:56 > wget

=> `lathe-modelling1.pdf' Resolving greggspen.googlepages.com... 64.233.179.91, 64.233.179.93,

64.233.179.95 Connecting to greggspen.googlepages.com[64.233.179.91]:80... connected. HTTP request sent, awaiting response... 200 OK Length: 2,179,534 [application/octet-stream] 100%[====================================================================================================================>] 2,179,534 153.18K/s ETA 00:00

20:47:00 (161.62 KB/s) - `lathe-modelling1.pdf' saved [2179534/2179534] ======================================================================

Only one try -- and a *lot* faster than your connections (37X). If you have a slow connection, it *might* be that the server is timing out on you, or it might be problems with the link somewhere between you and the server.

Do you have traceroute (or a Windows equivalent) available to you? It might be interesting to see what that says -- including the times. Here is what I see at the moment:

====================================================================== traceroute to googlepages.l.google.com (64.233.179.91), 30 hops max, 40 byte packets 1 Sponge (10.0.0.50) 1 ms 0 ms 0 ms 2 d-and-d.com (72.244.230.1) 1 ms 1 ms 1 ms 3 172.31.255.253 (172.31.255.253) 18 ms 24 ms 20 ms 4 192.168.22.65 (192.168.22.65) 156 ms 185 ms 231 ms 5 ge-5-1-189.ipcolo2.Washington1.Level3.net (63.210.31.9) 10 ms 9 ms 10 ms 6 ae-21-52.car1.Washington1.Level3.net (4.68.121.50) 15 ms 14 ms 15 ms 7 GOOGLE-INC.car1.Level3.net (4.79.228.38) 15 ms 9 ms 15 ms 8 72.14.238.136 (72.14.238.136) 25 ms 66.249.95.149 (66.249.95.149) 24 ms

24 ms 9 72.14.239.17 (72.14.239.17) 25 ms 209.85.248.129 (209.85.248.129) 25 ms 72.14.238.153 (72.14.238.153) 24 ms 10 72.14.238.178 (72.14.238.178) 30 ms 35 ms 35 ms 11 hs-in-f91.google.com (64.233.179.91) 25 ms 23 ms 25 ms ======================================================================

Both the number of hops needed and the times at each hop could have a bearing on your problems.

Good Luck, DoN.

And some are from Model Engineer, and may still be under copyright. After the last (?) change of ownership, many of the older less popular plans were passed on to the X Plans group.

Steve R.

Cool stuff! Now I'll experiment a bit. Thanks again!

S.

Yabbut... *we* were the ones in the know. Then (as now), it's up to us to do the correcting. ;-)

I don't think so. Since I've been using Linux, we've lived in... oh, let's see here... NM, FL, NH, NM, FL, AZ, NH, MN, AZ, NM, NH. In case you're interested, the full list is here:

Pretty much in the '90s, people said it right, and then as Windows users started using it, they started saying it with a short i.

When I learned to read, there were rules for pronouncing things. Granted, there were as many exceptions as rules, but there were rules, nonetheless.

I'd probably put it another way. Even Linus gave up on correctiong people, and got to the point where he shrugged and said, "I don't care how you pronounce it."

Personal preference I guess.

--20:46:45-- => `lathe-modelling1.pdf'

2,179,534 153.18K/s ETA 00:00 72.14.238.153 (72.14.238.153) 24 ms

Hi Don,

I'm sure the connection speed has a lot to do with it, which I already commented on. I'm on dial-up and usually have around a 48k connection.

Traceroute does reveal some convolutions:

Both your Wget log and Steve's show much faster connections than I can ever hope to achieve :)

I'll almost bet that everyone else having problems with this site are also using slow connections. The server is giving up on us too soon.

No big deal to me, that is why I keep things like Wget around. Problem solving.

Thanks for the log/info. I have often wondered if similar problems like this were speed related or not (shrug).