SoylentNews

I got an email earlier today saying that a 15 dollar credit was being applied to my Microsoft Account. When I logged in no such credit had been added to my account. I sent a "reply to:" the source of the email "billing@microsoft.com". It bounced fast.

While poking around in my account info I discovered that I could not add an email account I possess. It was listed as unavailable but I don't know how that could be since I only have one MS account.

So I went looking for a support email address. There is no such thing. There is "Chat" = 21 minute wait. There is "Phone" = 4 minute wait. Now all I need is simple information which is not listed on the web site. First is why did they send me an email that is meaningless or a possible phishing scam? and secondly why can't I add my own email address as an alias to my MS account? Neither of these things are of significant enout to invest 21 minutes to "chat" with some one or "4 minutes" for a phone call...

Microsoft why don't you innovate and actually make it easy to provide info and solve problems? Instead all you've done for me is wasted my time and made me angry because you don't believe in email. Alas...

Some may have been following my "playing" with Markov filters. [See: Markov's Wisdom] In reflecting on the output of the content of the "filter text", it appears to me that there may be a kind of "fractal dimension" built into language which is further refined by an individual's unique "voice" and "hearing" when communicating.

Why we marvel at the pseudo-communication of a filtered output is that the boundary between comprehensible and incomprehensible gets repeatedly gets crossed. What is happening is that the Markov filter showing "inconclusively" the boundary between communication and a stream of unstructured words. Sometimes the recombination of words creates something new and interesting-- which is why we appreciate it. One thing to remember is that the "newness" is not in the message itself but in how we have ourselves "filtered" the textual stream [see (e) below].

Understand I'm not a professional communication theorist. I'm a general specialist who knows a little bit about many things and a lot about nothing and possesses little or no real background in the theory, mathematics, or coding of information.

So forgive me while I think out loud for a moment and display my ignorance:

One might summarize Claude Shannon's information theory as: A text source contains a valid message [i.e. is comprehensible] when at least three things are present: (a) the components of the language contain "words" which represents abstract meaning of both tangible and intangible "things". (b) The message is composed in a specific language which is defined by its grammar and (c) the message is appropriately constructed in that grammar.

The meaning of the words, the language and its grammar defines a "dimensional boundary" along which a message "travel" to be considered comprehensible. Each language / grammar dimensional boundary is unique, hence translation between languages may have difficulty if there are no "words / grammar" which 'map' between languages. [An aside: One may wonder if we will actually understand the message if SETI were to discover a signal. Is their language and /or mathematics comprehensible to a human recipient?]

The dimensional boundary noted above might be considered the "entropy" of the message. One might also consider it as a kind of "fractal dimension" which the collection of words and letters must "fall along" to be comprehensible.

A further refinement of the expression in the textual content of the message comes from (d) the intellect and emotional state of the user of the language which composes the message which effects the composition of the message. This results in a unique assemblage and changes the message entropy [or dimension]. Sometimes this changes the entropy for better or worse. One might consider the best writers and orators [coders too!] convey their message with the least amount of information loss.

The receiver of the message must also possess a significant enough understanding of (a),(b),(c) and (d) to perceive if the message has reached a "dimension" to be comprehensible. Given that the recipient also has (e) their own intellectual and emotional state which effects the 'reception" of the message-- what may be a clear message at the source may change when it is "filtered" by the recipient.

The wonder here is that humans [and possibly other species] possess an "analog" information processor that can "comprehend on the fly" the entropy [dimension] of a message yet we have not been able to understand the process well enough to actually create an artificial language processor [in other words an AI] which possesses (a,b,c,d,e) above. We may understand that a message with a 'dimension' of "R" is comprehensible we have not yet discovered why this is so. Nor have we yet understood how that "R" might relate to messages of "S", "T", "U", "V", or the other dimensional boundaries which lie along the same coastline are related [or not]

I suspect that there may be a way to determine what the informational boundary for a set of works [books, essays, oratory, code] enough that a "processor" could be created to simulate the output of the "analog" creator. It remains to be seen if the next step could be taken to create a processor that crosses the line between being a "mimic" of a textual stream to becoming an actual creator of a communication stream.

If one can cross that boundary then it can be asked: If we assume that the boundary dimension of "R" represents a specific ability to emit output of information.... "Does a specific boundary dimension represent fully in a deterministic way the full ability of its possessor?

Which is to say can we make an AI smarter just by changing "boundary dimension"? Yes I realize that boundary dimension is created by the interaction of a possibly large set of variables... which brings on another one of those things to think about: what are those variables?

We can roughly measure the coastline of Britain and determine the dimension of that coastline. Can we do the same for language?

I'd appreciate your comments.

Dissociate Press is a Markov filter implimented as a Perl module which is based based upon an EMACs script as the Man page says:

This module provides the function `dissociate', which implements a Dissociated Press algorithm, well known to Emacs users as "meta-x dissociate". The algorithm here is by no means a straight port of Emacs's `dissociate.el', but is instead merely inspired by it.

The Book of Proverbs has a worthy ancestor in the baked brick tablets which contain the Sumerian Proverbs. A good sampling of these can be found in The Electronic Text Corpus of Sumerian Literature

I thought I'd give a sample of what Sumerian Proverbs look like after being "filtered" with the Dissociate Perl Script. If you are wondering the quotes were "cherry picked" but were taken verbatim from the output stream. I verified that none of the quotes below appeared in exactly the same form in the original input stream. Some punctuation has been added or removed but the "output streams were not tampered with.

An example: How lowly is the poor? It is the food basket.

You can find this at Markov's Wisdom

Last night I decided to fire up FreeBasic to do a quick file hack of converting folder of files each with a list of surnames in a standard ASCII text file into a file which can be used in a Random Generator app named TableSmith. For ease of use I thought I'd use something like the following algorithm:

Open filenameA which contains a list of filenames to be munged
Do Until eof(filenameA)

Read a line from the filenameA list and assign it to filenameB
open filenameB for input
create filenameC based upon modified filenameB
open output filenameC

Do Until eof(filenameb)

Input Line filenameB, somestring$
Write FilenameB, SomeString$

loop

Loop

It did not work. FreeBasic's string handling does not appear to work the way I thought it did. I tried many ways of doing this and I tripped over the undocumented features of String Arrays in freebasic. What I ended up with were "fragments of each element of filenameB".

I'm not sure if its the documentation, the compiler, or just a "1d10T error between the keyboard and chair". So I'm putting FB to the side and going to try to do this in freepascal... if all else fails I'll fall back on TP7 for DOS... where I know I can get it to work.

This is what I get for letting my skills get rusty.

Why do I want to use these antiquated languages? a 1d10T error: Because Perl, Python, and most OO languages are not 1d10T brain readable.

This ain't the geezer your looking for. Move along.

PS: formatting code in HTML in a S/N journal is time consuming / difficult. Too bad there is no WYSIWYG or even an [indent] [\indent]. I had to use block quotes to get a rough format that was readable.

for a casual programmer?

Perl and Python are not good languages to use for a "casual programmer" such as myself. My reasons are as follows:

Their learning curve is steep.

They require an external "runtime" and /or libraries.

They cannot "compile" a stand-alone executable [using wrappers makes the process too complex]

What are the alternatives?

"Scripting" Languages have many limitations. I've used TCL/Tk, Perl, Python, JavaScript. Of these TCL/Tk is the easiest to use simply because TCL is pretty straight forward [within its limitations] and Tk is integrated with TCL so building user facing windowing apps was fairly simple. None of the other languages noted above are as "easy" to write in as they were written for purposes other than human readability or easy functionality. "Back in the day" I used the IBM Basic compiler [MS basic compiler], Turbo Pascal, Realizer [AKA CA-Realizer was a Win31 / Win95 basic compiler that Computer Associates bought and then let MS kill http://en.wikipedia.org/wiki/CA-Realizer ] to write actual "production release" software. More recently I have used both FreeBasic and FreePascal.

I'm one of those foolish oldtimer's that think that OO languages add needless complexity while not actually providing added value to Rapid App Development. So talking C++, Java [and variants] and other OO languages at me is pretty pointless.

JavaScript is said to be the "coming thing" except it seems that it suffers from dialect fragmentation / implementation issues [Anyone remember how MS broke things with its own version of Java?] and last I looked it also had "math issues" [i.e. math is not done by simple expression "A=15+6" there is no specific number types-- Number is too general]

Many years ago [1998?] I wrote an app with JavaScript to run in the original Netscape browser that read in data, processed it and then wrote to disk the results. It was a bitch to write. I did it despite poorly documented features [disk reads and writes at the time were "supposed to be sandboxed". It turns out they were not. I am unconvinced that JS is actually a way forward since it was not designed to be a "general purpose" language.

I'd love to hear what others think the best general purpose language is for a "casual programmer".

Recently I looked in my "old code repository" to see if I had the source for "A Travesty Generator for Micros" from Byte Magazine [November 1984]. I did not find the code for the Byte version of Travesty but I did find a couple of old DOS apps: Markv and "Babble!".

MarkV is a markov chain text twister written in C. The most famous of the Markv versions was an IRC / USENET bot that emitted posts and comments. For an example of what MarkV can do to a list of "wisdom quotes" see http://markovswisdom.blogspot.com/2014/02/the-book-of-markovs-reminders.html

"Babble!" was a closed source DOS "shareware" app from the 1990s that allowed multiple Markov chains from a maximum of 4 text source travesty streams to be mixed into a single output stream. An example of this can be found here: http://markovswisdom.blogspot.com/2014/02/markov-and-shakespeare-reveal-wisdom-of.html The output combines a small snippet of the "wisdom quotes" + "Hamlet" + Jabberwocky" + "Romeo and Juliette". The problem with the version of "Babble!" I have is a) it is closed source and b) the shareware is crippleware in that the text streams are very limited. I'd very much like to have or develop a useable version of this.

Unfortunately there do not seem to be too many "modern" open source "travesty generators" available and those that I have found are not written in programming languages [Perl, Python] that I am a fluent programmer [which is to say I find neither Perl nor Python are not casual programmer friendly... but we'll save that for another journal entry]

One other thing I have noticed is that there do not seem to be too many "modern" implementations using Markov Chains and other methods to twist text available.

Anyone know where I can lay my hands on source code for "text twister" or Travesty generators?