Re: What Makes a substance Bizarre?

[jmfbahciv]

Peter Flass wrote:
> On 4/29/2013 9:02 AM, Shmuel (Seymour J.) Metz wrote:
>> In <klf213$70u$1...@dont-email.me>, on 04/26/2013
>> at 07:29 PM, Peter Flass <Peter...@Yahoo.com> said:
>>
>>> I support shipping arms to the Syrian rebels,
>>
>> Including AQ?
>>
>> FWIW, I support providing air cover *if* the Arab League asks us too
>> and someone is willing to provide nearby basing. I won't hold my
>> breathe; there seem to be diplomatic or policial issues for every one
>> of the neighboring countries.
>>
>
> That's why I think we need to imbed some "contractors." They could get
> to know the players and make judgements as to who to support and who to
> avoid. Naturally there'd be some leakage, but AQ has plenty of access
> to weapons now. If we boosted the folks we want to support it might
> help to de-incentivize the people we don't.
>
But AQ will be the leaders after everything is done. The most viscious
will retain political power while the others just want to be left
alone and get on with their lives.

/BAH

[jmfbahciv]

Dan Espen wrote:
> jmfbahciv <See....@aol.com> writes:
>
>> Dan Espen wrote:
>>> jmfbahciv <See....@aol.com> writes:

<snip>

>>> http://en.wikipedia.org/wiki/Gastarbeiter
>>>
>>> Seriously, why don't you just type things into Google?
>>> There's a whole trove of information out there.
>>
>> When I need to use Google, I go to the public library.
>
> I saw in your response to the Rod thing that you say using the web
> breaks your computer?
>
> That's weird.


Not only that but it's fucking annoying.

>
> And a damn shame. The Wikipedia thread on guest workers in Germany
> is a bit of interesting trivia.


OK. I'll try to remember to look it up when I get to library again.
>
> I'm not a mac user, but I'm guessing someone here might be able to help
> if you explained what happens.
>
I don't remember the details. I'm pretty sure it has to do with Java
(apps?) and the fact that I turned popups off. With them on, it
would take more than a day to down load one web site.

/BAH

[jmfbahciv]

Walter Bushell wrote:
> In article <PM0004DB7...@ac817e21.ipt.aol.com>,
> jmfbahciv <See....@aol.com> wrote:
>
>> John Levine wrote:
>> >>With databases and spreadsheets there is less need for a standalone
>> >>sort program as in the mainframe world.
>> >
>> > Less, but definitely not none.
>> >
>> > I have a daily job that downloads the .COM zone file, extracts the
>> > nameservers and such from it, and updates a local database with the
>> > changes. The only way I can make it reasonably fast is to sort the
>> > list of nameservers and do it like a merge, with one sequential pass
>> > through the database.
>> >
>> > There are 250 million nameservers in the file, so if I did it the
>> > obvious way by checking each record in the file against the database,
>> > it would take far more than 24 hours for each daily run.
>> >
>> but..but...but...today's system are so fast!!! ;-)
>>
>> Just another fact demonstrating that, if we provide an infinite
>> resource, the computer biz will use it up in 5 months.
>>
>> /BAH
>
> Not if you charge enough for the use of the resource. I remember using
> Compuserve and it was fun, until I got the bill.
>
that's why the PC biz took off.

/BAH

[jmfbahciv]

Charles Richmond wrote:
> "Peter Flass" <Peter...@Yahoo.com> wrote in message
> news:klllu2$5f9$2...@dont-email.me...
>> On 4/28/2013 6:30 PM, Charlie Gibbs wrote:
>>> In article <icy5c2i...@home.home>, des...@verizon.net (Dan Espen)
>>> writes:
>>>
>>>> I never worked on a commercial system that didn't have a vendor
>>>> supplied sort.
>>>
>>> I don't think I did either. But on one system I did have to write
>>> a little program that tested whether the file to be sorted was empty
>>> so I could bypass the sort - because this particular sort utility
>>> regarded an empty file as an error and aborted the job.
>>>
>>
>> Dummy record?
>>
>
> Dummy programmer! ;-)

nah, the software didn't have END= or EOF= options in the
READ statements. That was common in software which grew up
reading cards. Moving to disk files was traumatic for some
packages and lanugages.

/BAH

[jmfbahciv]

Rod Speed wrote:
>
>
> "jmfbahciv" <See....@aol.com> wrote in message
> news:PM0004DB7...@ac817e21.ipt.aol.com...
>> Rod Speed wrote:
>>>
>>>
>>> "jmfbahciv" <See....@aol.com> wrote in message
>>> news:PM0004DB6...@aca2d61b.ipt.aol.com...
>>>> Walter Bushell wrote:
>>>>> In article <icli84j...@home.home>, Dan Espen <des...@verizon.net>
>>>>> wrote:
>>>>>
>>>>>> Programmers don't write sort logic, they use the languages sort verb.
>>>>>> If you catch someone writing a sort, you send them back where they
>>>>>> came
>>>>>> from.
>>>>>
>>>>> Somebody has to write the sort routines. I remember my boss writing a
>>>>> sort for the Trash-80 which didn't come with a sort.
>>>
>>>> What did people do in the aulden days
>>>> when there weren't any sort packages?
>>>
>>> Write one.
>>
>> Now reread the thread.
>
> No point, that is what happened when there wasn’t one.
>
>
You do have trouble putting yourself in someone else's shoes
to understand their viewpoiint. Dan has an opinion on the
subject matter and it comes from his work experience. IOW,
he didn't write sorts nor did his coworkers.

/BAH

[jmfbahciv]

Walter Banks wrote:
>
>
> jmfbahciv wrote:
>
>> My Dad went to a one-room school which taught 1-8.
>> He ended up learning things above his age. I would
>> have liked that for arithmetic.
>
> That was a definite advantage of a one room school
> learning stuff above your grade. The curriculum was
> also organized to help the teachers. Many subjects
> material was taught on alternate years so that when
> it was taught in the classroom it was taught to two
> grades. Social studies for example common to
> grades 3 and 4.
>
> Hard to believe the changes, school had a barn
> to house the horses during the day. For recreation
> in some winters we set up a trap line to catch
> wild rabbits, mink and weasel for their fur. The catch
> yielded a few dollars each at the end of the winter.
> I think we got paid 90 cents for a mink skin.

Dad talked about checking his trap lines going to/from
school. He also carried a gun for shooting dinner or
varmits and sometimes a fishing pole when Grandma
wanted fish for supper.

/BAH

[jmfbahciv]

Dan Espen wrote:
> jmfbahciv <See....@aol.com> writes:
>
>> Dan Espen wrote:
>>> jmfbahciv <See....@aol.com> writes:

<snip>

>> And they are all web-based. Pineaplle was the only one I could
>> find which didn't require an internal knowledge of unix and
>> was not web-based.
>
> Google says there are a bunch of readers.
>
> Of course I'd recommend GNUS (Emacs based).
> No "internal knowledge of Unix" required.
>
> But there are whole bunch of for pay and free newsreaders.
>
> Reeder
> NetNewsWire
> Cream
> Pulp
> Gruml
> Cappuccino
> ...
>


Some of these are new. Cappuccino and NNW were web-based, IIRC.
You went to their web site to pull up the stuff you wanted. Taht
would take too long. My baud rates are usually around 37000.

/BAH

[jmfbahciv]

hanc...@bbs.cpcn.com wrote:
> On Apr 29, 8:50 am, jmfbahciv <See.ab...@aol.com> wrote:
>
>> Sounds like you never used the marvelous input and output
>> procedure options of the sort verb.
>
> Yes, I have, and based on experience, I believe a separate sort step
> makes the _overall process_ more maintainable as opposed to having a
> more complex program with the input and output procedures.

You must have had very different congruent data than what I dealt
with.

/BAH

[jmfbahciv]

Rod Speed wrote:
>
>
> "jmfbahciv" <See....@aol.com> wrote in message
> news:PM0004DB7...@ac817e21.ipt.aol.com...

<snip>

>>> Seriously, why don't you just type things into Google?
>>> There's a whole trove of information out there.
>>
>> When I need to use Google, I go to the public library.
>

> Utterly insane.
>
> Even you should be able to get another FREE discarded
> laptop that you could use if you are too stupid to reload
> the one you use for usenet.

The wain for one page to download is measured in hours, not seconds
and it never completes.

/BAH

[jmfbahciv]

Scott Lurndal wrote:
> Dan Espen <des...@verizon.net> writes:

>>jmfbahciv <See....@aol.com> writes:
>
>>> And they are all web-based. Pineaplle was the only one I could
>>> find which didn't require an internal knowledge of unix and
>>> was not web-based.
>>
>>Google says there are a bunch of readers.
>>
>>Of course I'd recommend GNUS (Emacs based).
>>No "internal knowledge of Unix" required.
>>
>>But there are whole bunch of for pay and free newsreaders.
>>
>>Reeder
>>NetNewsWire
>>Cream
>>Pulp
>>Gruml
>>Cappuccino
>>...
>

> Not sure what she means by "web-based", unless she's confusing it with
GUI-based.

You have to download pages from the site of the newsreader everytime you
invoke the newsreader. The one I used before was Seamonkey. It
got really bad waiting for the shite to download just so I could read/write
ASCII.

>
> For text-only, tin is available for MAC.


I looked at that and read the instructions. It required more sophisticated
knowledge of interneals than I had. Of course, it didn't proved that
knowledge nor a pointer to it.

>
> For GUI, pan is a good one.

It offends me that I have to execute so many instructions just to push/pull
ASCII.

/BAH

[jmfbahciv]

Peter Flass wrote:
> On 4/29/2013 8:50 AM, jmfbahciv wrote:
>> Peter Flass wrote:
>>> On 4/28/2013 5:46 PM, Dan Espen wrote:
>>>> Walter Bushell <pr...@panix.com> writes:
>>>>
>>>>> In article <ic7gjmk...@home.home>, Dan Espen <des...@verizon.net>

>>>>> wrote:
>>>>>
>>>>>> Seriously, why don't you just type things into Google?
>>>>>> There's a whole trove of information out there.
>>>>>

>>>>> Rhetorical question much?
>>>>
>>>> Being an eternal optimist, just a helpful suggestion.
>>>>
>>>
>>> It's gotten so I hate to be separated from my iPad. It's just so simple
>>> to look things up: Where have I seen this actor/actress before; where is
>>> this quote from; where the heck is Chechnya; etc.
>>>
>> How much does all that cost? I don't have a cell phone either ;-)
>>
>
> As I recall you don't have high-speed internet either,

Right. So I end up waiting for downloads only to observe that things are
fucked up with Javaese mess.

>which would be a
> problem. If you have cable or DSL all you need is a wi-fi router, and
> some services may have one included, otherwise <$100. Then you need the
> tablet. I have an iPad, but android tablets are around $200, AFAIK.
> Best investment I ever made.


The ads I get from ATT and the cable company talk about prices way above $100
and a long-term contract.

/BAH

[Shmuel Metz]

In <PM0004DB7...@ac817e21.ipt.aol.com>, on 04/29/2013

at 12:50 PM, jmfbahciv <See....@aol.com> said:

>And they are all web-based.

Nonsense.

>Pineaplle was the only one I could find which didn't require an
>internal knowledge of unix and was not web-based.

Plenty of newws clients are neither web based not require an internal
knowledge of Unix. In fact, some are only available on windoze.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to spam...@library.lspace.org

[Scott Lurndal]

"Bill Leary" <Bill_...@msn.com> writes:
>"Peter Flass" wrote in message news:klm8le$hbb$4...@dont-email.me...
>> On 4/29/2013 10:05 AM, Scott Lurndal wrote:
>>> Besides
>>> self-modifying code (i.e. what we call to JIT code generation),
>>
>> Interesting statement. These days everyone is taught that self-modifying
>> code = evil, while JIT = good.
>
>Not just "these days." I've been programming for close to forty years and
>it's always been, if not exactly "evil," at least "bad practice."
>Exceptions, of course, but pretty rare and generally due to the requirements
>of the instruction set. Such as a machine where the port number for I/O was
>part of the instruction but you wanted the handler to work with two or more
>similar devices. I saw this a few times in the minicomputer days with
>serial ports and interrupt handlers.
>
>> I never really connected the dots to see that they're the same thing.
>
>No dots. They're not the same thing. JIT is delaying code generation until
>just before its execution. Self modifying code is a program changing itself
>while it's running.

A JIT will dynamically generate code that will be executed immediately. That has
the exact same effects on caches as self-modifying code. Take Hotspot for
example: it is _exactly_ a "program changing itself while it's running".


scott

[Dan Espen]

Java apps need an okay to run.
Java scripts usually aren't that big.
Blocking pop-ups is a good thing and won't contribute to slowness.

On dial up, you might want to block images,
you definitely want to block flash,
and you would want to block ads.

All of that should be easy to do but you'll have to download plugins.

Using Google consists of typing a search term into the search bar or
going to www.google.com. All of that should be fast, even on dial up.


For some of us, dial up seems like using an IBM 402 to run our business
in 2013,

I just upgraded my FIOS from 25Mbps to 75Mbps.

--
Dan Espen

[Dan Espen]

I used Pan for a while.
What I really liked was the simple one character short cuts.
Just like using a text only app.

--
Dan Espen

[Dan Espen]

One thing to bear in mind.
When you switch news readers or news servers the reader may consider
all the posts on the server unread. You don't want to even try to
download all the headers in that case. Just mark everything read
and only read new posts after that.

--
Dan Espen

[Walter Bushell]

In article <nfhvn8plng8g9gblh...@4ax.com>,
Nick Spalding <spal...@iol.ie> wrote:

> Peter Flass wrote, in <kloc1e$vdq$1...@dont-email.me>
> on Tue, 30 Apr 2013 08:15:11 -0400:

>
> > On 4/29/2013 8:15 PM, Bill Leary wrote:
> > > "Peter Flass" wrote in message news:klm8le$hbb$4...@dont-email.me...
> > >> On 4/29/2013 10:05 AM, Scott Lurndal wrote:
> > >>> Besides
> > >>> self-modifying code (i.e. what we call to JIT code generation),
> > >>
> > >> Interesting statement. These days everyone is taught that
> > >> self-modifying code = evil, while JIT = good.
> > >
> > > Not just "these days." I've been programming for close to forty years
> > > and it's always been, if not exactly "evil," at least "bad practice."
> > > Exceptions, of course, but pretty rare and generally due to the
> > > requirements of the instruction set. Such as a machine where the port
> > > number for I/O was part of the instruction but you wanted the handler to
> > > work with two or more similar devices. I saw this a few times in the
> > > minicomputer days with serial ports and interrupt handlers.
> >

> > It was pretty common in my (360) DOS experience to have a routine branch
> > to some one-time initialization code, and the initialization change that
> > to a noop, or for a program to activate branches based on the contents
> > of a control card. Once I started writing OS code I wanted to make
> > everything reentrant, so I stopped coding this way. I think it was a
> > very common technique on 1401s, etc.
>
> In a 1401 without the Store-B-Register instruction the only way you
> could set up a return from a subroutine was to plug the address to be
> returned to into the exit branch of the subroutine before entering the
> subroutine.

There were machines that the subroutine call stored the address of the
calling instruction in the called location and jumped to the next
location. Arguments (or their addresses) were stored after the calling
instruction.

The older compilers in the DEC 10 series used this technique IIRC,
for the 20s this had to be replaced by loading a register with the
return addy, because you wanted the code segments to be reentrant and
thus usable by many jobs at once. Ah, that series had an excellent
machine language and the assembler was just about a high level
language.

--
Gambling with Other People's Money is the meth of the fiscal industry.
me -- in the spirit of Karl and Groucho Marx

[hanc...@bbs.cpcn.com]

On Apr 30, 3:17 am, "Charlie Gibbs" <cgi...@kltpzyxm.invalid> wrote:

> My faith in computer vendors' software offerings was never that great,
> and it evaporated completely when I started writing utilities that
> ran rings around Univac's stuff.  (When I was young and idealistic
> I donated some of my own stuff - they dumbed it down and filed off
> the serial numbers, but copied a few sentences word for word into
> their own documentation and inherited a few subtle bugs that could
> have been used as fingerprints.)

Having used both Univac (90 series) and IBM machines, I felt Univac's
software (eg compilers, monitors, etc) left a lot to be desired
compared to IBM.

[Walter Bushell]

In article <D2Qft.292401$r%2.6...@fed13.iad>,

sc...@slp53.sl.home (Scott Lurndal) wrote:

> A JIT will dynamically generate code that will be executed immediately.
> That has
> the exact same effects on caches as self-modifying code. Take Hotspot for
> example: it is _exactly_ a "program changing itself while it's running".
>
>
> scott

The rest of us weren't thinking of the effects on caches, but instead
thinking in terms of program maintainability.

[hanc...@bbs.cpcn.com]

On Apr 30, 9:32 am, jmfbahciv <See.ab...@aol.com> wrote:

> I heard a news report today that potato chips are getting caffeine
> added?  Is caffeine becoming as prevalent as high fructose corn
> syrup?

http://www.cbsnews.com/8301-204_162-57582007/fda-probing-caffeinated-products-from-gum-to-chips/

[hanc...@bbs.cpcn.com]

On Apr 30, 9:32 am, jmfbahciv <See.ab...@aol.com> wrote:

> Don't forget about all the TMP files SORTs write while sorting.  That
> is one of the reasons for an INPUT procedure...you can eliminate
> records to be sorted.  Once upon a time, computer systems didn't have
> that kind of disk space except for the huge disk farms which IBM
> hard/software knew how to manage.

A standalone sort has the INCLUDE and OMIT options to select records.
This works very fast, faster than an executed COBOL program selecting
records. Further, records may be reformatted, so long records can be
condensed.

[Scott Lurndal]

Walter Bushell <pr...@panix.com> writes:
>In article <D2Qft.292401$r%2.6...@fed13.iad>,
> sc...@slp53.sl.home (Scott Lurndal) wrote:
>
>> A JIT will dynamically generate code that will be executed immediately.
>> That has
>> the exact same effects on caches as self-modifying code. Take Hotspot for
>> example: it is _exactly_ a "program changing itself while it's running".
>>
>>
>> scott
>
>The rest of us weren't thinking of the effects on caches, but instead
>thinking in terms of program maintainability.

Which still applies. Try debugging a hotspot failure :-).

scott

[Bill Leary]

"Scott Lurndal" wrote in message news:D2Qft.292401$r%2.6...@fed13.iad...
>"Bill Leary" <Bill_...@msn.com> writes:
>> ((..omitted..))

>>> No dots. They're not the same thing. JIT is delaying code generation
>>> until just before its execution. Self modifying code is a program
>>> changing itself while it's running.
>
> A JIT will dynamically generate code that will be executed immediately.

Yes.

> That has the exact same effects on caches as self-modifying code.

The cache is not relevant to the issue. If you're going to use this
argument then everything, including data, is "self modifying code."

> Take Hotspot for example: it is _exactly_ a "program changing itself
> while it's running".

No. The program was not changed while running. The executable of the
program came into existence just before it was executed. All that's
happened is that (1) the moment of executable generation has changed from
some longer period before it was to be executed to immediately before
execution and (2) if you run the program again, it does it again. It is not
modifying itself.

I'd assumed you were misunderstanding what JIT meant, but now I'm thinking
you don't understand what self-modifying code is. Perhaps through never
having encountered it?

- Bill

[Dan Espen]

Faster than COBOL is a stretch.

How can sort be any faster than

IF PART-NO-SUFFIX EQUAL '1'

Please, save me from arcane SORT control card record selection without
benefit of copy books.

No generating control cards is not the same thing.

--
Dan Espen

[Bill Leary]

"Peter Flass" wrote in message news:kloc1e$vdq$1...@dont-email.me...

>On 4/29/2013 8:15 PM, Bill Leary wrote:

>> ((..omitted..))

>> Not just "these days." I've been programming for close to forty years
>> and it's always been, if not exactly "evil," at least "bad practice."
>> Exceptions, of course, but pretty rare and generally due to the
>> requirements of the instruction set. Such as a machine where the port
>> number for I/O was part of the instruction but you wanted the handler to
>> work with two or more similar devices. I saw this a few times in the
>> minicomputer days with serial ports and interrupt handlers.
>
> It was pretty common in my (360) DOS experience to have a routine branch
> to some one-time initialization code, and the initialization change that
> to a noop,

I've never seen that one, but I can see where that could be an efficiency
argument to self-modify.

> or for a program to activate branches based on the contents of a control
> card.

I'm not sure I understand this example.

> Once I started writing OS code I wanted to make everything reentrant, so
> I stopped coding this way. I think it was a very common technique on
> 1401s, etc.

Thanks for a bit of insight. Those machines were just before I got into
programming.

- Bill

[Bill Leary]

"Walter Bushell" wrote in message
news:proto-6BD350....@news.panix.com...

> In article <nfhvn8plng8g9gblh...@4ax.com>,
> Nick Spalding <spal...@iol.ie> wrote:

> ((..omitted..))

>> In a 1401 without the Store-B-Register instruction the only way you
>> could set up a return from a subroutine was to plug the address to be
>> returned to into the exit branch of the subroutine before entering the
>> subroutine.
>
> There were machines that the subroutine call stored the address of the
> calling instruction in the called location and jumped to the next
> location. Arguments (or their addresses) were stored after the calling
> instruction.

I've worked with that, but I didn't think of it as self modifying code. I
saw it as a storage location followed by code. That is, the instructions
were not being modified, just the location set aside to store the return
address. Never the less, I can see your point that what was being thought
of as code space was modified, even if no instructions were.

- Bill

[hanc...@bbs.cpcn.com]

On Apr 30, 11:57 am, Dan Espen <des...@verizon.net> wrote:

> > A standalone sort has the INCLUDE and OMIT options to select records.
> > This works very fast, faster than an executed COBOL program selecting
> > records.  Further, records may be reformatted, so long records can be
> > condensed.
>
> Faster than COBOL is a stretch.
> How can sort be any faster than
> IF PART-NO-SUFFIX EQUAL '1'
> Please, save me from arcane SORT control card record selection without
> benefit of copy books.
> No generating control cards is not the same thing.

Admittedly, I haven't done a benchmark comparison (like I did with the
COMP stuff), but the sort steps do run very fast, and they seem to run
faster than selecting/excluding records under program control. As
someone mentioned, the sort process is heavilly optimized, and that
would include INCLUDEs or other control statements as well as the I/
O. And yes, admittedly coding the INCLUDE statements is tedious since
you have to refer to file position and get the syntax exactly right.

[hanc...@bbs.cpcn.com]

On Apr 30, 12:01 pm, "Bill Leary" <Bill_Le...@msn.com> wrote:

> > or for a program to activate branches based on the contents of a control
> > card.
>
> I'm not sure I understand this example.

An assembly programmer showed me an application program, where by the
use of clever binary and the registers, he was able to set up the
appropriate branch far faster than we would do it in COBOL.

But as mentioned, that was back in the days of 64k machines and very
slow CPUs as compared to today. Unless the process was _extremely_
large and complex, it wouldn't save significant throughput time to
make it worthwhile.

[Dan Espen]

I did a test last year.

Overall the job elapsed time was halved.

What I really like though is that no one can mess with anything,
no multiple steps, no control cards, no temporary files, no restart
instructions, just submit the job and the report comes out.

--
Dan Espen

[Scott Lurndal]

"Bill Leary" <Bill_...@msn.com> writes:
>"Scott Lurndal" wrote in message news:D2Qft.292401$r%2.6...@fed13.iad...
>>"Bill Leary" <Bill_...@msn.com> writes:
>>> ((..omitted..))
>>>> No dots. They're not the same thing. JIT is delaying code generation
>>>> until just before its execution. Self modifying code is a program
>>>> changing itself while it's running.
>>
>> A JIT will dynamically generate code that will be executed immediately.
>
>Yes.
>
>> That has the exact same effects on caches as self-modifying code.
>
>The cache is not relevant to the issue. If you're going to use this
>argument then everything, including data, is "self modifying code."

Actually, from the hardware perspective, the cache is the -only- thing
relevent.

Consider a harvard architecture with split I/D caches. Many processors
don't bother to snoop the I-cache on D-cache or DMA reads, which means
that self-modifying code causes significant consistency problems. This applies whether
its code modifying a branch address in the instruction stream or
whether hotspot is generating JIT code for some bytecode sequence that will
be executed immediately.

>

>> Take Hotspot for example: it is _exactly_ a "program changing itself
>> while it's running".
>
>No. The program was not changed while running. The executable of the
>program came into existence just before it was executed. All that's
>happened is that (1) the moment of executable generation has changed from
>some longer period before it was to be executed to immediately before
>execution and (2) if you run the program again, it does it again. It is not
>modifying itself.

Sure it is. It's altering the instruction stream of the executable to
execute dynamically generated code. And sometimes even modifying the
generated code later.

The original VMware would replace code sequences in the operating systems
it was managing (linux or windows) to run-time paravirtualize the guest
OS to avoid problematic instruction sequences (e.g. SLDT/SGDT, et. alia).

I'd consider that also to be self-modifying code.

>
>I'd assumed you were misunderstanding what JIT meant, but now I'm thinking
>you don't understand what self-modifying code is. Perhaps through never
>having encountered it?

I've written it, particularly on PDP-8 (one can do some clever things updating
operate bits in the instruction word) and on burroughs mainframes.

I consider JIT to be no different from self-modifying code from the hardware
perspective, as it is progressively and incrementally generating code to
replace hot bytecode sequences on the fly over a relatively large period of
time and executed _in the context of the generator_.

I consider any runtime generated machine code, executed by the thread/process/core
generating the code to be self-modifying
(including PLT updates in libc), which may be a broader definition than yours.

[Stan Barr]

On Tue, 30 Apr 2013 08:24:43 +0100, Ahem A Rivet's Shot
<ste...@eircom.net> wrote:

> On Mon, 29 Apr 2013 13:05:09 -0400
> Peter Flass <Peter...@Yahoo.com> wrote:
>
>> On 4/29/2013 10:05 AM, Scott Lurndal wrote:
>> > Besides
>> > self-modifying code (i.e. what we call to JIT code generation),
>>
>> Interesting statement. These days everyone is taught that

>> self-modifying code = evil, while JIT = good. I never really connected

>> the dots to see that they're the same thing.
>

> They're not the same thing at all. JIT is just in time compilation,
> something I first saw in BASIC for the Newbrain when we called it a
> semi-compiler. Instead of interpreted BASIC the Newbrain compiled one line
> at a time as they were encountered and threw away compiled code when memory
> was getting tight. It gave the 'brain wonderful benchmark scores compared to
> interpreted BASICs because most BASIC benchmarks of the day were tight
> loops which got compiled first time round.
>
> AIUI JIT code generation is the same trick applied to things like
> Java bytecode interpreters.
>

Most Forth systems did/do the same thing. The system I'm still using keeps
everything as source and compiles to direct threaded code at load time.
A fast one pass compiler keeps speed up.

--
Cheers,
Stan Barr plan.b .at. dsl .dot. pipex .dot. com

The future was never like this!

[sdrat]

"Walter Bushell" <pr...@panix.com> wrote in message
news:proto-A9CE35....@news.panix.com...
> In article <au7rk6...@mid.individual.net>,
> "sdrat" <sd...@gfr.com> wrote:
>
>> "Peter Flass" <Peter...@Yahoo.com> wrote in message
>> news:klllq8$5f9$1...@dont-email.me...

>> > On 4/28/2013 5:46 PM, Dan Espen wrote:

>> >> Walter Bushell <pr...@panix.com> writes:
>> >>
>> >>> In article <ic7gjmk...@home.home>, Dan Espen <des...@verizon.net>
>> >>> wrote:
>> >>>

>> >>>> Seriously, why don't you just type things into Google?
>> >>>> There's a whole trove of information out there.
>> >>>

>> >>> Rhetorical question much?
>> >>
>> >> Being an eternal optimist, just a helpful suggestion.
>> >>
>> >
>> > It's gotten so I hate to be separated from my iPad. It's just so
>> > simple
>> > to look things up: Where have I seen this actor/actress before; where
>> > is
>> > this quote from; where the heck is Chechnya; etc.
>>

>> Yeah, I've had the PC next to me when even watching TV for decades now.
>>
>> A desktop tho, not an ipad or tablet. I prefer that format, even for
>> ebook
>> reading.
>
> My laptop is my TV. Plug in tuner via USB.

Screen's too small for me.

I currently use one of the PC 24" widescreen monitors for the
TV but that�s because my glasses have always been optimised
for the PC screens and I'm years late in updating them, so the
massive great TV isn't ideal for reading subtitles. Due to update
the glasses in a couple of weeks.

[sdrat]

"Peter Flass" <Peter...@Yahoo.com> wrote in message

news:klobh2$sbu$1...@dont-email.me...
> On 4/29/2013 9:02 AM, Shmuel (Seymour J.) Metz wrote:
>> In <klf213$70u$1...@dont-email.me>, on 04/26/2013
>> at 07:29 PM, Peter Flass <Peter...@Yahoo.com> said:
>>
>>> I support shipping arms to the Syrian rebels,
>>
>> Including AQ?
>>
>> FWIW, I support providing air cover *if* the Arab League asks us too
>> and someone is willing to provide nearby basing. I won't hold my
>> breathe; there seem to be diplomatic or policial issues for every one
>> of the neighboring countries.
>>
>
> That's why I think we need to imbed some "contractors." They could get to
> know the players and make judgements as to who to support and who to
> avoid. Naturally there'd be some leakage, but AQ has plenty of access to
> weapons now. If we boosted the folks we want to support it might help to
> de-incentivize the people we don't.

Much more likely to end up with the same farce the US ended up
with when it supported Saddam during the Iran/Iraq war instead.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message

news:PM0004DB9...@aca2316d.ipt.aol.com...

> Shmuel (Seymour J.) Metz wrote:

>> In <proto-7A1A3C....@news.panix.com>, on 04/26/2013
>> at 05:08 PM, Walter Bushell <pr...@panix.com> said:
>>
>>>Alcohol and nicotine are some of the hardest drugs out there,
>>>certainly harder than Maryjane. Alcohol withdrawal can kill and
>>>tobacco is apparently the most addictive of drugs, except perhaps
>>>for coffee and sugar.
>>
>> I used to be a heavy coffee drinker, but when I was really busy ar
>> work I'd forget to get up nad buy coffee; no withdrawal effects.
>> OTOH, a 36-hour day did affect me, with or without coffee.

> I heard a news report today that potato chips are getting caffeine added?

Just because its on a news report doesn’t make it gospel.

> Is caffeine becoming as prevalent as high fructose corn syrup?

Nope, and nothing like it either.

I haven't bothered with caffeine for more than a couple of decades now.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004DB9...@aca2316d.ipt.aol.com...

> Dan Espen wrote:
>> jmfbahciv <See....@aol.com> writes:
>>
>>> Dan Espen wrote:
>>>> jmfbahciv <See....@aol.com> writes:
>
> <snip>
>
>>>> http://en.wikipedia.org/wiki/Gastarbeiter
>>>>

>>>> Seriously, why don't you just type things into Google?
>>>> There's a whole trove of information out there.
>>>

>>> When I need to use Google, I go to the public library.
>>

>> I saw in your response to the Rod thing that you say using the web
>> breaks your computer?
>>
>> That's weird.

> Not only that but it's fucking annoying.

But completely trivial to fix, just reload the laptop.

Not a shred of rocket science whatever required.

>> And a damn shame. The Wikipedia thread on guest
>> workers in Germany is a bit of interesting trivia.

> OK. I'll try to remember to look it up when I get to library again.

And that’s another thing you should automate, the
list of stuff to do next time you are at the library.

>> I'm not a mac user, but I'm guessing someone here
>> might be able to help if you explained what happens.

> I don't remember the details. I'm pretty sure it has to do with
> Java (apps?) and the fact that I turned popups off. With them
> on, it would take more than a day to down load one web site.

Oh bullshit.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004DB9...@aca2316d.ipt.aol.com...

More fool you.

What matters is how useable the total is, including web browsing.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004DB9...@aca2316d.ipt.aol.com...

> Rod Speed wrote:
>>
>>
>> "jmfbahciv" <See....@aol.com> wrote in message

>> news:PM0004DB7...@ac817e21.ipt.aol.com...

>>> Rod Speed wrote:
>>>>
>>>>
>>>> "jmfbahciv" <See....@aol.com> wrote in message

>>>> news:PM0004DB6...@aca2d61b.ipt.aol.com...
>>>>> Walter Bushell wrote:
>>>>>> In article <icli84j...@home.home>, Dan Espen <des...@verizon.net>
>>>>>> wrote:
>>>>>>
>>>>>>> Programmers don't write sort logic, they use the languages sort
>>>>>>> verb.
>>>>>>> If you catch someone writing a sort, you send them back where they
>>>>>>> came
>>>>>>> from.
>>>>>>
>>>>>> Somebody has to write the sort routines. I remember my boss writing a
>>>>>> sort for the Trash-80 which didn't come with a sort.
>>>>
>>>>> What did people do in the aulden days
>>>>> when there weren't any sort packages?
>>>>
>>>> Write one.
>>>
>>> Now reread the thread.
>>
>> No point, that is what happened when there wasn’t one.

> You do have trouble putting yourself in someone
> else's shoes to understand their viewpoiint.

Bullshit. I wrote a number of sorts, because DEC didn’t provide any.

> Dan has an opinion on the subject matter
> and it comes from his work experience. IOW,
> he didn't write sorts nor did his coworkers.

But what was being discussed was what
was done when there was no sort available.

Its just not viable to do nothing, so you had to write one.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004DB9...@aca2316d.ipt.aol.com...

> Dan Espen wrote:
>> jmfbahciv <See....@aol.com> writes:
>>
>>> Dan Espen wrote:
>>>> jmfbahciv <See....@aol.com> writes:
>
> <snip>
>

>>> And they are all web-based. Pineaplle was the only one I could
>>> find which didn't require an internal knowledge of unix and
>>> was not web-based.
>>
>> Google says there are a bunch of readers.
>>
>> Of course I'd recommend GNUS (Emacs based).
>> No "internal knowledge of Unix" required.
>>
>> But there are whole bunch of for pay and free newsreaders.
>>
>> Reeder
>> NetNewsWire
>> Cream
>> Pulp
>> Gruml
>> Cappuccino
>> ...
>>
>
>

> Some of these are new. Cappuccino and NNW were web-based, IIRC.
> You went to their web site to pull up the stuff you wanted. Taht
> would take too long. My baud rates are usually around 37000.

I did it for years at that speed, works fine.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004DB9...@aca2316d.ipt.aol.com...

> Peter Flass wrote:
>> On 4/29/2013 9:02 AM, Shmuel (Seymour J.) Metz wrote:
>>> In <klf213$70u$1...@dont-email.me>, on 04/26/2013
>>> at 07:29 PM, Peter Flass <Peter...@Yahoo.com> said:
>>>
>>>> I support shipping arms to the Syrian rebels,
>>>
>>> Including AQ?
>>>
>>> FWIW, I support providing air cover *if* the Arab League asks us too
>>> and someone is willing to provide nearby basing. I won't hold my
>>> breathe; there seem to be diplomatic or policial issues for every one
>>> of the neighboring countries.
>>>
>>
>> That's why I think we need to imbed some "contractors." They could get
>> to know the players and make judgements as to who to support and who to
>> avoid. Naturally there'd be some leakage, but AQ has plenty of access
>> to weapons now. If we boosted the folks we want to support it might
>> help to de-incentivize the people we don't.

> But AQ will be the leaders after everything is done.

Nope, they have never been able to lead a damned thing.

> The most viscious will retain political power while the
> others just want to be left alone and get on with their lives.

It isn't just the most vicious that retain political power.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004DB9...@aca2316d.ipt.aol.com...

> Rod Speed wrote:
>>
>>
>> "jmfbahciv" <See....@aol.com> wrote in message

>> news:PM0004DB7...@ac817e21.ipt.aol.com...
>
> <snip>

>
>>>> Seriously, why don't you just type things into Google?
>>>> There's a whole trove of information out there.
>>>
>>> When I need to use Google, I go to the public library.
>>

>> Utterly insane.
>>
>> Even you should be able to get another FREE discarded
>> laptop that you could use if you are too stupid to reload
>> the one you use for usenet.
>
> The wain for one page to download is measured in hours, not seconds
> and it never completes.

Then there is something utterly fucked about that laptop.

I used those baud rates and a lot lower too for years, worked fine.

[Rod Speed]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004DB9...@aca2316d.ipt.aol.com...

> Peter Flass wrote:
>> On 4/29/2013 8:50 AM, jmfbahciv wrote:

>>> Peter Flass wrote:
>>>> On 4/28/2013 5:46 PM, Dan Espen wrote:
>>>>> Walter Bushell <pr...@panix.com> writes:
>>>>>

>>>>>> In article <ic7gjmk...@home.home>, Dan Espen <des...@verizon.net>

>>>>>> wrote:
>>>>>>
>>>>>>> Seriously, why don't you just type things into Google?
>>>>>>> There's a whole trove of information out there.
>>>>>>

>>>>>> Rhetorical question much?
>>>>>
>>>>> Being an eternal optimist, just a helpful suggestion.
>>>>>
>>>>
>>>> It's gotten so I hate to be separated from my iPad. It's just so
>>>> simple
>>>> to look things up: Where have I seen this actor/actress before; where
>>>> is
>>>> this quote from; where the heck is Chechnya; etc.
>>>>

>>> How much does all that cost? I don't have a cell phone either ;-)
>>>
>>
>> As I recall you don't have high-speed internet either,

> Right. So I end up waiting for downloads only to
> observe that things are fucked up with Javaese mess.

wikipedia doesn’t. Neither does google.

>>which would be a
>> problem. If you have cable or DSL all you need is a wi-fi router, and
>> some services may have one included, otherwise <$100. Then you need the
>> tablet. I have an iPad, but android tablets are around $200, AFAIK.
>> Best investment I ever made.

> The ads I get from ATT and the cable company talk
> about prices way above $100 and a long-term contract.

Bullshit.

[Rod Speed]

"Scott Lurndal" <sc...@slp53.sl.home> wrote in message

news:D2Qft.292401$r%2.6...@fed13.iad...
> "Bill Leary" <Bill_...@msn.com> writes:

>>"Peter Flass" wrote in message news:klm8le$hbb$4...@dont-email.me...

>>> On 4/29/2013 10:05 AM, Scott Lurndal wrote:
>>>> Besides
>>>> self-modifying code (i.e. what we call to JIT code generation),
>>>
>>> Interesting statement. These days everyone is taught that
>>> self-modifying
>>> code = evil, while JIT = good.
>>

>>Not just "these days." I've been programming for close to forty years and
>>it's always been, if not exactly "evil," at least "bad practice."
>>Exceptions, of course, but pretty rare and generally due to the
>>requirements
>>of the instruction set. Such as a machine where the port number for I/O
>>was
>>part of the instruction but you wanted the handler to work with two or
>>more
>>similar devices. I saw this a few times in the minicomputer days with
>>serial ports and interrupt handlers.
>>

>>> I never really connected the dots to see that they're the same thing.
>>

>>No dots. They're not the same thing. JIT is delaying code generation
>>until
>>just before its execution. Self modifying code is a program changing
>>itself
>>while it's running.

> A JIT will dynamically generate code that will be executed immediately.

> That has the exact same effects on caches as self-modifying code.

But not on the maintainability of the code.

> Take Hotspot for example: it is _exactly_ a
> "program changing itself while it's running".

But not on the maintainability of the code.

[Rod Speed]

"Scott Lurndal" <sc...@slp53.sl.home> wrote in message

news:u3Rft.90797$aK1....@fed02.iad...

Nope.

> Try debugging a hotspot failure :-).

Different matter entirely to how readable the code is.

[Rod Speed]

"Scott Lurndal" <sc...@slp53.sl.home> wrote in message

news:svSft.206302$LZ6.1...@fed14.iad...

But not relevant to the reason real self modifying code was
frowned upon, it can be more difficult to read and maintain.

[Dave Garland]

On 4/29/2013 5:27 PM, JimP. wrote:
> On 29 Apr 2013 20:03:02 GMT, greymausg <ma...@mail.com> wrote:

>> On 2013-04-29, Peter Flass <Peter...@Yahoo.com> wrote:
>>> On 4/28/2013 5:46 PM, Dan Espen wrote:

>>>> Walter Bushell <pr...@panix.com> writes:
>>>>
>>>>> In article <ic7gjmk...@home.home>, Dan Espen <des...@verizon.net>
>>>>> wrote:
>>>>>
>>>>>> Seriously, why don't you just type things into Google?
>>>>>> There's a whole trove of information out there.
>>>>>
>>>>> Rhetorical question much?
>>>>
>>>> Being an eternal optimist, just a helpful suggestion.
>>>>
>>>
>>> It's gotten so I hate to be separated from my iPad. It's just so simple
>>> to look things up: Where have I seen this actor/actress before; where is
>>> this quote from; where the heck is Chechnya; etc.
>>>
>>
>>

>> Chechnya?... Tis slightly down from Berlin, sneakily renamed the Czech
>> Republic. The Capital, even sneakierly put on the map as Praha, really
>> Grozny.
>
> Nope. Chechnya is a former associate of Soviet Union. Which is why
> Putin doesn't like them. They sent a group to take over a Moscow movie
> theater a few years ago. The Russian Army pumped a sleep gas in that
> killed many.

..woosh..

[Andrew Swallow]

With JIT the source code remains unmodified. The machine code is
deleted from core at the end of the run. A new compile is performed the
next time the program runs.

On computers with I and D memory caches the machine code has to be
copied from the D cache to I cache for every new line of code.

Andrew Swallow

[Patrick Scheible]

jmfbahciv <See....@aol.com> writes:

> Dan Espen wrote:
>> jmfbahciv <See....@aol.com> writes:
>>
>>> Dan Espen wrote:
>>>> jmfbahciv <See....@aol.com> writes:
>
> <snip>
>
>>> And they are all web-based. Pineaplle was the only one I could
>>> find which didn't require an internal knowledge of unix and
>>> was not web-based.
>>
>> Google says there are a bunch of readers.
>>
>> Of course I'd recommend GNUS (Emacs based).
>> No "internal knowledge of Unix" required.
>>
>> But there are whole bunch of for pay and free newsreaders.
>>
>> Reeder
>> NetNewsWire
>> Cream
>> Pulp
>> Gruml
>> Cappuccino
>> ...
>>
>
>
> Some of these are new. Cappuccino and NNW were web-based, IIRC.
> You went to their web site to pull up the stuff you wanted. Taht
> would take too long. My baud rates are usually around 37000.

I'd set up leafnode, then the only things that have to be put through
the thin pipe are the articles, not all the gui shite. Then you could
use whatever newreader you like on your own news spool.

However, that does take a little bit of system administration to set up
and some space on the hard drive for whatever news groups you read.
Hm. I am assuming that since leafnode runs on the BSDs and Linux there
must be a Mac version of it.

-- Patrick

[Patrick Scheible]

hanc...@bbs.cpcn.com writes:

> On Apr 30, 9:32 am, jmfbahciv <See.ab...@aol.com> wrote:
>

>> I heard a news report today that potato chips are getting caffeine

>> added?  Is caffeine becoming as prevalent as high fructose corn
>> syrup?
>
> http://www.cbsnews.com/8301-204_162-57582007/fda-probing-caffeinated-products-from-gum-to-chips/

Good lord. Like loading them up with salt and fat wasn't enough.

-- Patrick

[Patrick Scheible]

hanc...@bbs.cpcn.com writes:

> On Apr 29, 6:08 pm, Dave Garland <dave.garl...@wizinfo.com> wrote:
>
>> On 4/29/2013 3:46 PM, brad wrote:
>
>> > I would expect a
>> > turned off cell phone to stay charged for several months.
>
> Correct. But if the phone is left 'on' it will lose its charge in a
> few days or so even if it doesn't make or receive any calls.
>
>
>> ITYM "on".  Though it's not the receiver per se that's the problem,
>> it's that the phone (transmitter) periodically checks in with the
>> tower so the network knows where to send calls.
>
> It does not do this when the phone is turned 'off'. When a phone is
> off, calls go to voice mail, which is handled by the central office,
> not the phone.

Doesn't it put an indicator on the screen showing that you have voice
mail or a missed call, even if the phone is "off"?

Doesn't the phone keep in touch with what the nearest cell tower is, so
it doesn't have to hunt around when someone makes a call?

I'm relucant to disbelieve what the FBI agent tells me. He was most
emphatic and apparently traces people this way quite routinely.

-- Patrick

[JimP.]

On 30 Apr 2013 09:33:15 GMT, greymausg <ma...@mail.com> wrote:

>On 2013-04-29, JimP <pongb...@cableone.net> wrote:
>> On 29 Apr 2013 20:03:02 GMT, greymausg <ma...@mail.com> wrote:
>>>On 2013-04-29, Peter Flass <Peter...@Yahoo.com> wrote:

>>>> On 4/28/2013 5:46 PM, Dan Espen wrote:
>>>>> Walter Bushell <pr...@panix.com> writes:
>>>>>

>>>>>> In article <ic7gjmk...@home.home>, Dan Espen <des...@verizon.net>
>>>>>> wrote:
>>>>>>
>>>>>>> Seriously, why don't you just type things into Google?
>>>>>>> There's a whole trove of information out there.
>>>>>>
>>>>>> Rhetorical question much?
>>>>>
>>>>> Being an eternal optimist, just a helpful suggestion.
>>>>>
>>>>
>>>> It's gotten so I hate to be separated from my iPad. It's just so simple
>>>> to look things up: Where have I seen this actor/actress before; where is
>>>> this quote from; where the heck is Chechnya; etc.
>>>>
>>>
>>>
>>>Chechnya?... Tis slightly down from Berlin, sneakily renamed the Czech
>>>Republic. The Capital, even sneakierly put on the map as Praha, really
>>>Grozny.
>>
>> Nope. Chechnya is a former associate of Soviet Union. Which is why
>> Putin doesn't like them. They sent a group to take over a Moscow movie
>> theater a few years ago. The Russian Army pumped a sleep gas in that
>> killed many.
>> .
>

>Wow, so Sarah Palin is _wrong_? :)

Is she ever correct ?

>Actually, the Chechens were an independent people until taken over by
>the Czarist Russians about 180 years ago, the resistance _then_ lasted
>for years. Their capital, Grozny was a main aim of the German attack
>in 1942, in which the Chechens are supposed to have helped, with the
>usual result. At the end of the main revolt, some years ago, more bombs,
>etc, had been used in Grozny than at Stalingrad. Coincidentally :), they
>control the (oil) Pipeline from azerbaijan.

Well, I was saying they were part of the SovUnion and wanted out of
post-Collapse Russia.
.
--
Brushing aside the thorns so I can see the stars.
http://www.linuxgazette.net/ Linux Gazette
http://dice.drivein-jim.net/ my dice collection
http://poetry.drivein-jim.net/ Aug 26, 2009

[sdrat]

"Patrick Scheible" <k...@zipcon.net> wrote in message
news:86ip33d...@chai.my.domain...

> hanc...@bbs.cpcn.com writes:
>
>> On Apr 29, 6:08 pm, Dave Garland <dave.garl...@wizinfo.com> wrote:
>>
>>> On 4/29/2013 3:46 PM, brad wrote:
>>
>>> > I would expect a
>>> > turned off cell phone to stay charged for several months.
>>
>> Correct. But if the phone is left 'on' it will lose its charge in a
>> few days or so even if it doesn't make or receive any calls.
>>
>>
>>> ITYM "on". Though it's not the receiver per se that's the problem,
>>> it's that the phone (transmitter) periodically checks in with the
>>> tower so the network knows where to send calls.
>>
>> It does not do this when the phone is turned 'off'. When a phone is
>> off, calls go to voice mail, which is handled by the central office,
>> not the phone.

> Doesn't it put an indicator on the screen showing that you
> have voice mail or a missed call, even if the phone is "off"?

Not even possible if the phone is turned off.

> Doesn't the phone keep in touch with what the nearest cell tower
> is, so it doesn't have to hunt around when someone makes a call?

Not when the phone is turned off, only when its still turned on.

> I'm relucant to disbelieve what the FBI agent tells me.

He either doesn’t have a fucking clue about the basics
or is deliberately lying to you for some reason.

> He was most emphatic and apparently
> traces people this way quite routinely.

Because most don’t turn their phone off.

[Peter Flass]

On 4/30/2013 8:43 AM, Walter Bushell wrote:
> In article <proto-CA6105....@news.panix.com>,
> Walter Bushell <pr...@panix.com> wrote:
>
>> In article <ouSdnUQdMNOPkOLM...@giganews.com>,

>> "Bill Leary" <Bill_...@msn.com> wrote:
>>
>>> Not just "these days." I've been programming for close to forty years and
>>> it's always been, if not exactly "evil," at least "bad practice."
>>> Exceptions, of course, but pretty rare and generally due to the
>>> requirements
>>> of the instruction set. Such as a machine where the port number for I/O
>>> was
>>> part of the instruction but you wanted the handler to work with two or more
>>> similar devices. I saw this a few times in the minicomputer days with
>>> serial ports and interrupt handlers.
>>

>> One machine I worked on had the number of bits in a shit or rotate in
>> the instruction itself. If you wanted a variable shift you could mask
>> in the number of bits to shift. Self modifying code but not too hard
>> to debug.
>
> s/shit/shift/
>

Probably right the first time -- ugly!!! If the count couldn't be in a
register they needed an EXECUTE instruction.

--
Pete

[Peter Flass]

Or I guess they expected you to code a shift/rotate of 1 bit and run it
in a loop.


--
Pete

[Peter Flass]

On 4/30/2013 9:32 AM, jmfbahciv wrote:
> Peter Flass wrote:
>> On 4/29/2013 8:50 AM, jmfbahciv wrote:

>>> Peter Flass wrote:
>>>> On 4/28/2013 5:46 PM, Dan Espen wrote:
>>>>> Walter Bushell <pr...@panix.com> writes:
>>>>>
>>>>>> In article <ic7gjmk...@home.home>, Dan Espen <des...@verizon.net>
>>>>>> wrote:
>>>>>>
>>>>>>> Seriously, why don't you just type things into Google?
>>>>>>> There's a whole trove of information out there.
>>>>>>
>>>>>> Rhetorical question much?
>>>>>
>>>>> Being an eternal optimist, just a helpful suggestion.
>>>>>
>>>>
>>>> It's gotten so I hate to be separated from my iPad. It's just so simple
>>>> to look things up: Where have I seen this actor/actress before; where is
>>>> this quote from; where the heck is Chechnya; etc.
>>>>

>>> How much does all that cost? I don't have a cell phone either ;-)
>>>
>>
>> As I recall you don't have high-speed internet either,
>
> Right. So I end up waiting for downloads only to observe that things are
> fucked up with Javaese mess.
>

>> which would be a
>> problem. If you have cable or DSL all you need is a wi-fi router, and
>> some services may have one included, otherwise <$100. Then you need the
>> tablet. I have an iPad, but android tablets are around $200, AFAIK.
>> Best investment I ever made.
>
>
> The ads I get from ATT and the cable company talk about prices way above $100
> and a long-term contract.
>

> /BAH
>

The $100 was for the wi-fi part, or are you talking about getting it as
a service from AT&T? Don't do it. I couldn't believe they were selling
to to unsuspecting victims - my neighbor asked me about it. You can buy
a Netgear or other box and just plug it in.

--
Pete

[Peter Flass]

On 4/30/2013 10:18 AM, Dan Espen wrote:
>
> Using Google consists of typing a search term into the search bar or
> going to www.google.com. All of that should be fast, even on dial up.
>

That's what got me to switch from Yahoo to Google years ago. Yahoo was
plastering their search screen with all kinds of cruft and Google was
just a logo and a prompt.



--
Pete

[Peter Flass]

On 4/30/2013 10:29 AM, Walter Bushell wrote:
> In article <D2Qft.292401$r%2.6...@fed13.iad>,
> sc...@slp53.sl.home (Scott Lurndal) wrote:
>
>> A JIT will dynamically generate code that will be executed immediately.
>> That has
>> the exact same effects on caches as self-modifying code. Take Hotspot for
>> example: it is _exactly_ a "program changing itself while it's running".
>>
>>
>> scott
>
> The rest of us weren't thinking of the effects on caches, but instead
> thinking in terms of program maintainability.
>

I don't think self-modifying code is inherently any less maintainable
than any other. Hopefully you aren't modifying instructions randomly in
memory. Just comment an instruction that's modified and the
cross-reference will tell you all you need to know.

--
Pete

[Peter Flass]

On 4/30/2013 12:01 PM, Bill Leary wrote:
> "Peter Flass" wrote in message news:kloc1e$vdq$1...@dont-email.me...
>
>> or for a program to activate branches based on the contents of a
>> control card.
>
> I'm not sure I understand this example.

Suppose your program can print a number of different reports or generate
a number of different outputs from the same data. Read your control
card telling you which to print/write. following the read of a record
have a series of calls to the routines that generate each report, and
no-op the ones you don't want. That's a bit of a stretch, since it only
saves one instruction per report. I suppose you could use the same to
select fields to print, etc.


--
Pete

[Dan Espen]

AltaVista went the same way.

Wow, just visited AltaVista.
A really close imitation of Google.
Way too late.

--
Dan Espen

[Peter Flass]

On 4/30/2013 12:38 PM, Dan Espen wrote:
> hanc...@bbs.cpcn.com writes:
>

>> On Apr 30, 11:57 am, Dan Espen <des...@verizon.net> wrote:
>>
>>>> A standalone sort has the INCLUDE and OMIT options to select records.
>>>> This works very fast, faster than an executed COBOL program selecting
>>>> records. Further, records may be reformatted, so long records can be
>>>> condensed.
>>>
>>> Faster than COBOL is a stretch.
>>> How can sort be any faster than
>>> IF PART-NO-SUFFIX EQUAL '1'
>>> Please, save me from arcane SORT control card record selection without
>>> benefit of copy books.
>>> No generating control cards is not the same thing.
>>
>> Admittedly, I haven't done a benchmark comparison (like I did with the
>> COMP stuff), but the sort steps do run very fast, and they seem to run
>> faster than selecting/excluding records under program control. As
>> someone mentioned, the sort process is heavilly optimized, and that
>> would include INCLUDEs or other control statements as well as the I/
>> O. And yes, admittedly coding the INCLUDE statements is tedious since
>> you have to refer to file position and get the syntax exactly right.
>
> I did a test last year.
>
> Overall the job elapsed time was halved.

The speedup is obviously not from the compare (IF) code. The sort doing
the I/O is a lot more efficient than COBOL doing it - it probably does
multiple track reads and may well have timing tricks built in.

--
Pete

[Shmuel Metz]

In <1671.902T14...@kltpzyxm.invalid>, on 04/29/2013
at 11:07 PM, "Charlie Gibbs" <cgi...@kltpzyxm.invalid> said:

>Am I the only one who sees automatic "upgrades" as one of the
>biggest security holes that Microsoft (and others) has ever foisted
>off on an unsuspecting public?

I'd buy "one of the biggest", but they have some other real zingers.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to spam...@library.lspace.org

[Shmuel Metz]

In <kln5p8$10be$1...@leila.iecc.com>, on 04/30/2013
at 01:12 AM, John Levine <jo...@iecc.com> said:

>Knuth disagrees about the 1950s machines. The UNIVAC certainly had
>one, and poking around in bitsavers I find quite a few, such as the
>1956 one for the 705.

>It may be that the scientific machines were less likely to have a
>sort utility than the commercial ones.

Knuth says that there were sorts for, e.g., 650, 701, 702, Datamatic?
Please provide a quote. Note that the 650 was a fairly popular
commercial machine.

[Shmuel Metz]

In <klmqrh$9jv$1...@dont-email.me>, on 04/29/2013
at 05:08 PM, Dave Garland <dave.g...@wizinfo.com> said:

>Of course, people want to be able to receive calls when their phone
>is "off",

Not when the reason that their phone is off is to avoid disrupting,
e.g., a movie.

[Shmuel Metz]

In <klm883$hbb$2...@dont-email.me>, on 04/29/2013
at 12:57 PM, Peter Flass <Peter...@Yahoo.com> said:

>Plus all the job setup overhead in z/OS.

That's negligible for a large application. IMHO the decision should be
made based on what is most maintainable.

[Shmuel Metz]

In <PM0004DB9...@aca2316d.ipt.aol.com>, on 04/30/2013

at 01:32 PM, jmfbahciv <See....@aol.com> said:

>I heard a news report today that potato chips are getting caffeine
>added? Is caffeine becoming as prevalent as high fructose corn
>syrup?

In soft drinks, yes, but I never heard of it in potato chips before.

[Andrew Swallow]

There is a difference between a phone being turned off and a call being
ended. Once the call has been ended the phone will only transmit about
once every 20 minutes to when moved to a different area.

There is a different rule for land lines, the only way to switch one of
those off is to pull the wires out.

Andrew Swallow

[John Levine]

>>Knuth disagrees about the 1950s machines. The UNIVAC certainly had
>>one, and poking around in bitsavers I find quite a few, such as the
>>1956 one for the 705.
>
>>It may be that the scientific machines were less likely to have a
>>sort utility than the commercial ones.
>
>Knuth says that there were sorts for, e.g., 650, 701, 702, Datamatic?
>Please provide a quote. Note that the 650 was a fairly popular
>commercial machine.

No, he's talking about the history of sorting algorithms and notes
that they were well developed for the UNIVAC, and that sorting
algorithms were in the folklore throughout the 1950s. See Knuth vol
3, pages 386-388 for details.

Pugh et al in IBM's Early Computers describes the 702's sort program
on p. 345 and for the 705 on p. 346. Sort and the assembler were the
first programs written and tested on the 702. It mentions the sort
program for the 709 and 7090 in passing on p. 367.

I suppose it's possible there wasn't a sort program for the 650.
Sorting software wouldn't be very useful without tape drives to store
the data and it is my impression that 650s were typically card only.
If you're going to sort cards, you might as well use a card sorter
which was a lot cheaper than a computer.

There was a long article on sorting software in the JACM in 1956 but I
don't feel like paying $15 for it. If anyone's an ACM member, perhaps
they could download it and tell us what it says.

http://dl.acm.org/citation.cfm?id=320833

R's,
John
--
Regards,
John Levine, jo...@iecc.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. http://jl.ly

[hanc...@bbs.cpcn.com]

On Apr 30, 5:08 pm, Patrick Scheible <k...@zipcon.net> wrote:
> > It does not do this when the phone is turned 'off'.  When a phone is
> > off, calls go to voice mail, which is handled by the central office,
> > not the phone.
>
> Doesn't it put an indicator on the screen showing that you have voice
> mail or a missed call, even if the phone is "off"?

It does that after the phone is turned back on.

> Doesn't the phone keep in touch with what the nearest cell tower is, so
> it doesn't have to hunt around when someone makes a call?

It does that after the phone is turned back on. It continues to do
that--it has to--as the user moves around.

[hanc...@bbs.cpcn.com]

On Apr 30, 8:51 pm, John Levine <jo...@iecc.com> wrote:

> I suppose it's possible there wasn't a sort program for the 650.
> Sorting software wouldn't be very useful without tape drives to store
> the data and it is my impression that 650s were typically card only.
> If you're going to sort cards, you might as well use a card sorter
> which was a lot cheaper than a computer.

The 650 had a relatively long life. Initially it was card I/O with a
magnetic drum only, but later on they added various peripherals to it,
as well as some core memory. So, sorts were likely utilized for
machines that had adequate disks or tapes.

IIRC, it was originally intended for sci/eng use and had very limited
character ability, but many businesses found it useful. It turned out
to work well with table lookups as done by insurance companies.

[Dan Espen]

No doubt about it, IBM is facing competition and I believe both sorts
use every trick they can to run all out. (DFSORT and Syncsort).

If IBM figured out a faster way to read a file in DFSORT, one wonders
why they wouldn't use those same tricks for QSAM.

But using an internal sort removes two entire passes of the
input/output file. That's where the gain is.

I don't see how any "optimization" in SORT can make it do OMITs or
INCLUDEs faster than COBOL. In both cases, the program just has to
do a compare and branch.


--
Dan Espen

[Walter Bushell]

In article <klpkg5$g5e$2...@dont-email.me>,

EXECUTE DESIGNER IMMEDIATE ?

--
Gambling with Other People's Money is the meth of the fiscal industry.
me -- in the spirit of Karl and Groucho Marx

[sdrat]

"Andrew Swallow" <am.sw...@btinternet.com> wrote in message
news:4tKdnfypfdSi_B3M...@bt.com...

Yes, but we were discussing Dave turning his cell/mobile
phone off so that no one could track his movements.

> Once the call has been ended the phone will only transmit about once every
> 20 minutes to when moved to a different area.

Yes, but that was not what was being discussed.

> There is a different rule for land lines, the only way to switch one of
> those off is to pull the wires out.

Yes, we were discussing cell/mobile phones, not landlines.

By definition no one wants track where your landline phone is.

[Bill Leary]

"Peter Flass" wrote in message news:klplun$ncm$1...@dont-email.me...

> On 4/30/2013 12:01 PM, Bill Leary wrote:
>> "Peter Flass" wrote in message news:kloc1e$vdq$1...@dont-email.me...
>>
>>> or for a program to activate branches based on the contents of a
>>> control card.
>>
>> I'm not sure I understand this example.
>
> Suppose your program can print a number of different reports
> or generate a number of different outputs from the same data.
> Read your control card telling you which to print/write.

Ah, the lights just went on. I took "card" to mean a printed circuit with
components on it (i.e.: CPU card). You mean "card" as in "80 column punch
card."

The example makes sense now. Thanks.

- Bill

[Bill Leary]

"Peter Flass" wrote in message news:klpkj1$g5e$3...@dont-email.me...

> On 4/30/2013 7:45 PM, Peter Flass wrote:
>> On 4/30/2013 8:43 AM, Walter Bushell wrote:

>>> ((..omitted..))

>>>> One machine I worked on had the number of bits in a shit or rotate in
>>>> the instruction itself. If you wanted a variable shift you could mask
>>>> in the number of bits to shift. Self modifying code but not too hard
>>>> to debug.
>>>
>>> s/shit/shift/
>>>
>>
>> Probably right the first time -- ugly!!! If the count couldn't be in a
>> register they needed an EXECUTE instruction.
>>
>

> I guess they expected you to code a shift/rotate of 1 bit and run
> it in a loop.

The General Automation GA-16 had the XEC (eXEcute register Contents)
instruction. You'd build any one-word (sixteen bit) instruction in a
register, then execute it. The supposed reason for this instruction was
because you couldn't do self-modifying code when your program was in ROM.
If that's true, then it would seem at that time (the manual I've got is from
1979) self-modifying code didn't have quite the bad name it did later.

- Bill

[Shmuel Metz]

In <klpotn$lul$1...@leila.iecc.com>, on 05/01/2013

at 12:51 AM, John Levine <jo...@iecc.com> said:

>No, he's talking about the history of sorting algorithms

That has nothing to do with the availability of vwendor provided sort
software on 1950 vintage machines.

>It mentions the sort
>program for the 709 and 7090 in passing on p. 367.

As I recall the 709 came out shortly before the 7090, near the end of
the decade.

>I suppose it's possible there wasn't a sort program for the 650.
>Sorting software wouldn't be very useful without tape drives to
>store the data and it is my impression that 650s were typically card
>only.

While tape and disk drives were available, card only was far more
prevalent.

[John Levine]

>If that's true, then it would seem at that time (the manual I've got is from
>1979) self-modifying code didn't have quite the bad name it did later.

Self modifying code was the great insight of the Von Neumann machine
design, that a program could rewrite itself. Through the 1950s it not
only had a good name, it was pretty much impossible to write useful
programs without using it.

The reason we don't write programs that patch themselves any more is
that people invented indirect addressing and particularly index
registers. Most of the modifications in the code were to do
addressing, and index registers provide the same effect more
efficiently and without having to change the code.

Even though self modifying code is disreputable, modern computers do
it all the time. Program loading and shared libraries are
self-modifying code on a large and stylized scale.

[Dave Garland]

On 4/30/2013 5:16 PM, sdrat wrote:
>
>
> "Patrick Scheible" <k...@zipcon.net> wrote in message

>> I'm relucant to disbelieve what the FBI agent tells me.
>
> He either doesn’t have a fucking clue about the basics
> or is deliberately lying to you for some reason.
>
>> He was most emphatic and apparently
>> traces people this way quite routinely.
>
> Because most don’t turn their phone off.

I think that's the key. I hardly know anybody (except for me) who does.

A lot of people have their computer set to hibernate (or even "sleep")
when they press the "off" switch, too, and don't realize that there's
a difference. And my understanding is, that goes for a lot of
smartphones, too. The manufacturer doesn't want you turning it off
(to "enhance user experience"). The carrier certainly doesn't want
you turning it off. They may not make it easy, or obvious, how to do
it. (This is supposition on my part, I have very little experience
with smartphones.)

"But it turns on quicker that way."

I have heard that some smartphones do not in fact turn off without
taking the battery out. Dunno if it's true.

We're dealing with a public who (pre-GPS) had a hard time
understanding that when you called 911 (or whatever your emergency
number is), they didn't immediately know where you were, just like a
landline.

[Dave Garland]

On 4/30/2013 8:58 AM, Shmuel (Seymour J.) Metz wrote:
> In <klmqrh$9jv$1...@dont-email.me>, on 04/29/2013
> at 05:08 PM, Dave Garland <dave.g...@wizinfo.com> said:
>
>> Of course, people want to be able to receive calls when their phone
>> is "off",
>
> Not when the reason that their phone is off is to avoid disrupting,
> e.g., a movie.
>

Going by my exposure to them in movies/concerts/meetings, this is not
something that ever enters their mind.

They could, of course, set it to "vibrate" or something, they don't
actually need to turn it off. Every cellphone I've had, had some way
to mute the rings. But I don't think anyone ever does.

[Patrick Scheible]

You can unplug them from the phone network.

-- Patrick

[Andrew Swallow]

That how you pull the wires out.

[Rod Speed]

"Dave Garland" <dave.g...@wizinfo.com> wrote in message
news:klq2sm$bu3$1...@dont-email.me...

> On 4/30/2013 5:16 PM, sdrat wrote:
>>
>>
>> "Patrick Scheible" <k...@zipcon.net> wrote in message
>
>>> I'm relucant to disbelieve what the FBI agent tells me.
>>
>> He either doesn’t have a fucking clue about the basics
>> or is deliberately lying to you for some reason.
>>
>>> He was most emphatic and apparently
>>> traces people this way quite routinely.
>>
>> Because most don’t turn their phone off.
>
>
> I think that's the key. I hardly know anybody (except for me) who does.

I do too, primarily because I have to charge it much less often that way.

> A lot of people have their computer set to hibernate (or even "sleep")
> when they press the "off" switch, too, and don't realize that there's a
> difference. And my understanding is, that goes for a lot of smartphones,
> too. The manufacturer doesn't want you turning it off

I don’t buy that last. Its trivially easy to do with any phone.

The only thing they really do is make it hard to do accidentally for obvious
reasons.

> (to "enhance user experience").

Don’t buy that, its trivially easy to do.

> The carrier certainly doesn't want you turning it off.

Who gives a damn what the carrier wants ?

> They may not make it easy, or obvious, how to do it.

I don’t know of a single phone where it isnt easy and obvious.

> (This is supposition on my part, I have very little experience with
> smartphones.)

I have a wealth of experience with smartphones
and none of them make it hard to turn them off.

The most they ever do is make it hard to do that accidentally.

> "But it turns on quicker that way."

> I have heard that some smartphones do not in fact turn off without taking
> the battery out. Dunno if it's true.

Its bullshit and you can prove that by checking
how long the battery lasts with it turned off.

> We're dealing with a public who (pre-GPS) had a hard time understanding
> that when you called 911 (or whatever your emergency number is), they
> didn't immediately know where you were, just like a landline.

Different matter entirely. That’s not at all surprising.

[Rod Speed]

"Dave Garland" <dave.g...@wizinfo.com> wrote in message

news:klq3c4$drh$1...@dont-email.me...

Plenty do, particularly when it matters like when doing surgery etc.

Even kids in school do because their phone gets confiscated when it rings.

[ERSHC]

On Tue, 30 Apr 2013 18:15:50 -0700 (PDT), hanc...@bbs.cpcn.com
<hanc...@bbs.cpcn.com> wrote:
> On Apr 30, 5:08 pm, Patrick Scheible <k...@zipcon.net> wrote:
>> > It does not do this when the phone is turned 'off'.  When a phone is
>> > off, calls go to voice mail, which is handled by the central office,
>> > not the phone.
>>
>> Doesn't it put an indicator on the screen showing that you have voice
>> mail or a missed call, even if the phone is "off"?
>
> It does that after the phone is turned back on.
>
>

But how "off" is off? If the phone doesn't have a mechanical toggle
but uses a momentary contact switch to turn on/off (like mine does),
then in order for the system to "see" the "turn me on" push, it has
to be on at some level. On my phone (an HTC G1) the only way for it
to be all the way off is to pull the battery.

No, I don't bother to do that.

[Ibmekon]

On 30 Apr 2013 13:32:26 GMT, jmfbahciv <See....@aol.com> wrote:

>Dan Espen wrote:
>> jmfbahciv <See....@aol.com> writes:
>>
>>> Dan Espen wrote:
>>>> jmfbahciv <See....@aol.com> writes:
>
><snip>
>
>>> And they are all web-based. Pineaplle was the only one I could
>>> find which didn't require an internal knowledge of unix and
>>> was not web-based.
>>
>> Google says there are a bunch of readers.
>>
>> Of course I'd recommend GNUS (Emacs based).
>> No "internal knowledge of Unix" required.
>>
>> But there are whole bunch of for pay and free newsreaders.
>>
>> Reeder
>> NetNewsWire
>> Cream
>> Pulp
>> Gruml
>> Cappuccino
>> ...
>>
>
>
>Some of these are new. Cappuccino and NNW were web-based, IIRC.
>You went to their web site to pull up the stuff you wanted. Taht
>would take too long. My baud rates are usually around 37000.
>
>/BAH

At that rate I would soon get bored.

Someone here recently (a.f.c in the last year) - suggested using the
Windows HOSTS file of blocked addresses to stop adverts.
OK until a webpage just hangs on loading - and does not show at all.
I guess we will have to get used to incessant web brainwashing - like
the TV adverts.


Carl Goldsworthy

"All I ask is the chance to prove that money can't make me happy."
Spike Milligan

[Peter Flass]

On 4/30/2013 9:51 PM, Dan Espen wrote:
>
> No doubt about it, IBM is facing competition and I believe both sorts
> use every trick they can to run all out. (DFSORT and Syncsort).

As I recall there were other sort programs for IBM systems besides IBM
and Syncsort, but my memory fails me just now. Does anyone else recall any?

--
Pete

[Peter Flass]

On 5/1/2013 5:53 AM, Ibmekon wrote:
>
> Someone here recently (a.f.c in the last year) - suggested using the
> Windows HOSTS file of blocked addresses to stop adverts.
> OK until a webpage just hangs on loading - and does not show at all.
> I guess we will have to get used to incessant web brainwashing - like
> the TV adverts.

*That* is a huge annoyance. Often when a page is slow to load I see
that it's waiting on ad.doubleclick.net or similar. For dog's sake, if
you're going to shove advertising at me, at least make it quick to load...

--
Pete

[Ibmekon]

At a couple of companies we used CA-SORT from Computer Associates.

It could be configured by an input control deck to optimise for CPU
cycles or elapsed time.

I remember that it could dynamically load balance DASD devices
according to channels and devices available at run time.

The advice was too give it as many SORTWORK devices as available, it
would only allocate as needed.

Maybe it could be supplied with input selection crtieria - certainly I
never used such a feature.

The billing at these companies included monthly charges for CPU time
used.
The sort program was rented - and paid for itself in CPU time savings.

Also, reminds me of a sysprog - George Barnard.
He wrote his own version of IBM's ubiquitous file copy program
IEBGENER - his UNBGENER (UNB was in the company name).
His version clocked less than half (rough memory estimate) the IBM
version.
But did not suport any input parameters.
On the other hand , you could leave out LRECL and BLKSIZE , it then
performed a straight copy.
Mainly use for backup of course.

[Shmuel Metz]

In <klq28f$1gsl$1...@leila.iecc.com>, on 05/01/2013

at 03:30 AM, John Levine <jo...@iecc.com> said:

>Through the 1950s it not only had a good name, it was pretty much
>impossible to write useful programs without using it.

I believe that to this day self-modigying code is the norm in the LISP
community.

>Most of the modifications in the code were to do addressing, and
>index registers provide the same effect more efficiently and
>without having to change the code.

Often, but not always, even with a S/360 style Execute instruction.
Nonetheless, I had started avoiding nonrefreshable code by the late
1960's, except for machines with a return jump instruction.

[Ibmekon]

A trick that works sometimes, is to click 'X' to cancel the page load.
Immediately click to reload the current page.
Some html reads the cookie just stored and assumes the ad site is
unavailable.
Problem solved.
Simples !

[Shmuel Metz]

In <klquj8$2vb$1...@dont-email.me>, on 05/01/2013

at 07:44 AM, Peter Flass <Peter...@Yahoo.com> said:

>As I recall there were other sort programs for IBM systems besides
>IBM and Syncsort, but my memory fails me just now. Does anyone
>else recall any?

PLSORT, but there were others.

[Shmuel Metz]

In <klq3c4$drh$1...@dont-email.me>, on 04/30/2013

at 10:52 PM, Dave Garland <dave.g...@wizinfo.com> said:

>Going by my exposure to them in movies/concerts/meetings, this is
>not something that ever enters their mind.

Even after the "turn off your cell phone" announcement :-(

Still, the ones that do as they were asked turn them off in order to
not receive calls.

[Dan Espen]

I never encountered that kind of problem using the hosts file.
Of course, this wasn't Windows, it was Linux.

Far easier today is use an adblocker.
No hangs there either.

--
Dan Espen

[Dan Espen]

Ibmekon writes:

> On Wed, 01 May 2013 07:44:30 -0400, Peter Flass
> <Peter...@Yahoo.com> wrote:
>
>>On 4/30/2013 9:51 PM, Dan Espen wrote:
>>>
>>> No doubt about it, IBM is facing competition and I believe both sorts
>>> use every trick they can to run all out. (DFSORT and Syncsort).
>>
>>As I recall there were other sort programs for IBM systems besides IBM
>>and Syncsort, but my memory fails me just now. Does anyone else recall any?
>
> At a couple of companies we used CA-SORT from Computer Associates.
>
> It could be configured by an input control deck to optimise for CPU
> cycles or elapsed time.
>
> I remember that it could dynamically load balance DASD devices
> according to channels and devices available at run time.
>
> The advice was too give it as many SORTWORK devices as available, it
> would only allocate as needed.

The latest recommendation on SORTWK is don't supply anything.
The sort program will figure out what it needs just fine.

--
Dan Espen

[jmfbahciv]

John Levine wrote:
>>If that's true, then it would seem at that time (the manual I've got is from
>>1979) self-modifying code didn't have quite the bad name it did later.
>
> Self modifying code was the great insight of the Von Neumann machine
> design, that a program could rewrite itself. Through the 1950s it not
> only had a good name, it was pretty much impossible to write useful
> programs without using it.
>
> The reason we don't write programs that patch themselves any more is
> that people invented indirect addressing and particularly index
> registers. Most of the modifications in the code were to do
> addressing, and index registers provide the same effect more
> efficiently and without having to change the code.
>
> Even though self modifying code is disreputable, modern computers do
> it all the time. Program loading and shared libraries are
> self-modifying code on a large and stylized scale.
>

It's not really disreputable but it's a bitch to debug. In addtion,
on the PDP-10s, if a bit was changed by the code, it could no longer
be a sharable highseg. (There was a one-word exception but I'm not
going to get into that.)

In addition, the way dirty programming crooks got into other
sites' systems was to masquerade their back door or password
workaround by modifying innocuous code.

/BAH

[jmfbahciv]

The differnece could be the number of tmp files used.

/BAH

[jmfbahciv]

Rod Speed wrote:
>
>
> "jmfbahciv" <See....@aol.com> wrote in message
> news:PM0004DB9...@aca2316d.ipt.aol.com...

>> Dan Espen wrote:
>>> jmfbahciv <See....@aol.com> writes:
>>>
>>>> Dan Espen wrote:
>>>>> jmfbahciv <See....@aol.com> writes:
>>
>> <snip>
>>
>>>> And they are all web-based. Pineaplle was the only one I could
>>>> find which didn't require an internal knowledge of unix and
>>>> was not web-based.
>>>
>>> Google says there are a bunch of readers.
>>>
>>> Of course I'd recommend GNUS (Emacs based).
>>> No "internal knowledge of Unix" required.
>>>
>>> But there are whole bunch of for pay and free newsreaders.
>>>
>>> Reeder
>>> NetNewsWire
>>> Cream
>>> Pulp
>>> Gruml
>>> Cappuccino
>>> ...
>>>
>>
>>
>> Some of these are new. Cappuccino and NNW were web-based, IIRC.
>> You went to their web site to pull up the stuff you wanted. Taht
>> would take too long. My baud rates are usually around 37000.
>

> I did it for years at that speed, works fine.

Wrong tense...it worked fine. Now it won't with all the movies
one has to endure.

/BAH

>

[jmfbahciv]

Dan Espen wrote:
> jmfbahciv <See....@aol.com> writes:
>
>> Rod Speed wrote:
>>>
>>>
>>> "jmfbahciv" <See....@aol.com> wrote in message

>>> news:PM0004DB7...@ac817e21.ipt.aol.com...
>>
>> <snip>
>>
>>>>> Seriously, why don't you just type things into Google?
>>>>> There's a whole trove of information out there.
>>>>
>>>> When I need to use Google, I go to the public library.
>>>
>>> Utterly insane.
>>>
>>> Even you should be able to get another FREE discarded
>>> laptop that you could use if you are too stupid to reload
>>> the one you use for usenet.
>>
>> The wain for one page to download is measured in hours, not seconds
>> and it never completes.
>
> One thing to bear in mind.
> When you switch news readers or news servers the reader may consider
> all the posts on the server unread. You don't want to even try to
> download all the headers in that case. Just mark everything read
> and only read new posts after that.

I know that. I"ve switched three times. I'm talking about about
a web page where "a" implies one screen full.

/BAH

>

[jmfbahciv]

Patrick Scheible wrote:

> jmfbahciv <See....@aol.com> writes:
>
>> Dan Espen wrote:
>>> jmfbahciv <See....@aol.com> writes:
>>>
>>>> Dan Espen wrote:
>>>>> jmfbahciv <See....@aol.com> writes:
>>

>> <snip>
>>
>>>> And they are all web-based. Pineaplle was the only one I could
>>>> find which didn't require an internal knowledge of unix and
>>>> was not web-based.
>>>
>>> Google says there are a bunch of readers.
>>>
>>> Of course I'd recommend GNUS (Emacs based).
>>> No "internal knowledge of Unix" required.
>>>
>>> But there are whole bunch of for pay and free newsreaders.
>>>
>>> Reeder
>>> NetNewsWire
>>> Cream
>>> Pulp
>>> Gruml
>>> Cappuccino
>>> ...
>>>
>>
>>
>> Some of these are new. Cappuccino and NNW were web-based, IIRC.
>> You went to their web site to pull up the stuff you wanted. Taht
>> would take too long. My baud rates are usually around 37000.
>

> I'd set up leafnode, then the only things that have to be put through
> the thin pipe are the articles, not all the gui shite. Then you could
> use whatever newreader you like on your own news spool.
>
> However, that does take a little bit of system administration to set up
> and some space on the hard drive for whatever news groups you read.
> Hm. I am assuming that since leafnode runs on the BSDs and Linux there
> must be a Mac version of it.

That's always been one of my fallbacks if I ever needed to run my own
newsservice. however, sysadmin docs are still not as helpful as DEC's
manuals are. people still are assuming that Unix is run by a computer
center or on a single-user single-owner system.

/BAH

[jmfbahciv]

Peter Flass wrote:
> On 4/30/2013 9:32 AM, jmfbahciv wrote:
>> Peter Flass wrote:
>>> On 4/29/2013 8:50 AM, jmfbahciv wrote:
>>>> Peter Flass wrote:
>>>>> On 4/28/2013 5:46 PM, Dan Espen wrote:
>>>>>> Walter Bushell <pr...@panix.com> writes:
>>>>>>
>>>>>>> In article <ic7gjmk...@home.home>, Dan Espen <des...@verizon.net>

>>>>>>> wrote:
>>>>>>>
>>>>>>>> Seriously, why don't you just type things into Google?
>>>>>>>> There's a whole trove of information out there.
>>>>>>>

>>>>>>> Rhetorical question much?
>>>>>>
>>>>>> Being an eternal optimist, just a helpful suggestion.
>>>>>>
>>>>>
>>>>> It's gotten so I hate to be separated from my iPad. It's just so simple
>>>>> to look things up: Where have I seen this actor/actress before; where is
>>>>> this quote from; where the heck is Chechnya; etc.
>>>>>
>>>> How much does all that cost? I don't have a cell phone either ;-)
>>>>
>>>
>>> As I recall you don't have high-speed internet either,
>>
>> Right. So I end up waiting for downloads only to observe that things are
>> fucked up with Javaese mess.
>>
>>> which would be a
>>> problem. If you have cable or DSL all you need is a wi-fi router, and
>>> some services may have one included, otherwise <$100. Then you need the
>>> tablet. I have an iPad, but android tablets are around $200, AFAIK.
>>> Best investment I ever made.
>>
>>
>> The ads I get from ATT and the cable company talk about prices way above
$100
>> and a long-term contract.
>>
>> /BAH
>>
>
> The $100 was for the wi-fi part, or are you talking about getting it as
> a service from AT&T? Don't do it. I couldn't believe they were selling
> to to unsuspecting victims - my neighbor asked me about it. You can buy
> a Netgear or other box and just plug it in.

The pricing was for the service/month.

/BAH

>

[jmfbahciv]

Dan Espen wrote:
> jmfbahciv <See....@aol.com> writes:
>

<snip>

> For some of us, dial up seems like using an IBM 402 to run our business
> in 2013,

<grin> I still remember the joy I had when I got to use 300 BPS. And
I still marvel when I finally got a 2400 Baud line at work.

>
> I just upgraded my FIOS from 25Mbps to 75Mbps.
>
I got yet another sales brochure from ATT yesterday. The 25 costs
$40/month and the 75 costs over $70/month with a 2 year contract
and the fine print said there were additional costs.

But their selling point is that I get to save $1500/year. So when
they decide to start charging the regular prices, I'll get to be a
lot poorer. Do those marketeers really think that mentioning
how much the stuff really costs will sell more disservices?

/BAH

[jmfbahciv]

hanc...@bbs.cpcn.com wrote:
> On Apr 30, 9:32 am, jmfbahciv <See.ab...@aol.com> wrote:
>
>> Don't forget about all the TMP files SORTs write while sorting.  That
>> is one of the reasons for an INPUT procedure...you can eliminate
>> records to be sorted.  Once upon a time, computer systems didn't have
>> that kind of disk space except for the huge disk farms which IBM
>> hard/software knew how to manage.

>
> A standalone sort has the INCLUDE and OMIT options to select records.
> This works very fast, faster than an executed COBOL program selecting
> records. Further, records may be reformatted, so long records can be
> condensed.
>

Huh? The standalone sort I'm talking about doesn't have coding. We
may be talking about vastly different things.

/BAH