More Manuscript Recovery Apps

This blog introduces 2 more apps related to manuscript recovery, the PEI and the LEX. Finally a headline review.

Recovery Process

We are following an iterative design process to build out the tooling needed to work through the letter by letter recovery of the inspired text of the Bible. The first round of this iterative process is building out the data pipeline for that recovery process and the apps needed to display the work as it moves forward over the weeks to come.

The data pipeline is managing both the manuscript for that process and a parallel set of files that collectively form a lexicon. Because of the number of steps in the process, there are different apps for looking at various stages in the process. The apps themselves are linked off of our main paleo.in website.

The first of those apps was online a couple of weeks ago. The second app is now online too. The second app was also going to include a lexicon, but because of the size of that app, the lexicon is now its own app. Let me review the apps as they are currently running.

ASPI

The ASPI app went online a couple weeks ago. This app is the first in a series. It shows off the early stages in the data pipeline for the manuscript recovery. The job of the ASPI is to show the origin of the Paleo letter string used in the recovery process.

This is a simple goal, but it is still a big app with many parts that make it go.

The ASPI has 3 rows of content. The top row of text is in Aramaic script. The second row is in Syriac script. The 3rd row is in Paleo script with bubbled numbers for vowel points. That bottom row does its job, it shows off the letter string we are using for recovery.

That bottom row of Paleo across the ASPI becomes the top row in the first of this week's new apps. Let me introduce that app.

PEI

This app is also a standard Bible, currently with the standard OT and NT. The Apocrypha can be added later. This PEI app has standard navigation menus and our standard story headings to help readers move around and recognize where they are located.

There are currently 3 rows in this app. The top row is the Paleo string from the bottom of the ASPI. The second row is an indication as to the how well the first row's words pass audit. The third row is an English approximation for the Paleo words.

Let me mention a little more about each row in turn.

Top Row: Paleo

Unlike the ASPI, the PEI is standard left-to-right writing, which is the inspired writing direction, East to West, the movement of the sun, when viewed from the north.

Latin is the ancient witness to this writing direction. The conquering of Ai from the north is a textual parable about how the text and the recovery process only work when viewed from the north.

This top row is not exactly the same as the bottom row of the ASPI. Besides writing direction, there are 2 other changes. First, the bubbled numbers in the ASPI are not being shown here. This adds complexity to PEI display.

Analysis of the bubbles from the ASPI will show up in the Lexicon should it turn out to be interesting.

The second difference from the ASPI in the top row of the PEI is that the words found in the ASPI have been parsed into their constituent parts. This is a major first step in the recovery process.

This is an ongoing work effort, so not all words are yet parsed. Let me return to that parsing problem below.

Second Row: Audit Marks

The 2nd row provides an indication about how the word as shown is currently passing the audit system. These marks are subject to change. With the use of clever notation it is possible to pack a lot of data into that row.

Currently, the audit marks are displayed using a set of digits, from 0 through 5. Each digit marks the legal left positions for the word as given immediately above. No digits at all indicate the word as drawn above is not legal using the audit system.

These digits are a hint for our further work. I suspect the marks we are using will change as the work progresses.

Audit marks do not mean the word itself is known to be correctly spelled, nor does it mean the word connects to its neighbors in an audit passing way. Those details will come with time.

Third Row: English

Currently the app uses a caret, ^, in the English row to indicate that the Paleo word seen above in the top row does not yet have a parse entry in the lexicon. The English row uses a question mark, ?, to indicate there is no English entry for the otherwise parsed entry in the word above.

In the PEI app this week, only a very few words have some English typed in as a systems test. The English will be generated by software using the parsed paleo and the text of the BRB. I would guess full English is still a few weeks out. How this will be done is explained below.

LEX

The link here is to our other new app, a Paleo Lexicon. Like the PEI, it is linked off the front page of paleo.in, for anyone who wants to find it in the future.

BEWARE: This app is big, at only about 1/2 of expected content it rings in at 212MB. It takes noticable time to download. Let the blue line run to the right before messing around in the app.

If this gets to be a problem, I will drop the offline function, and let it serve pages on demand from the server. But, if I do that the app will not run offline. I will leave it as is for now. If anyone has trouble, please send me feedback.

When I was planning this work the Lexicon was going to be hung off the PEI as an appendix. But, the PEI is already a big app. There will be 1 page in the lexicon for each root. So an eventual page total of around 10,000 words is what I am expecting when this is finished.

If you have ever owned a printed Bible lexicon, say a Strongs Number reference book or an NIV lexicon, you can appreciate the size. Lexicons are many times larger than the Bible itself.

The lexicon app has 2 important sections. The first introduces the Paleo Alphabet. The second is organized around root words. Let me explain each section in turn.

Lex: Alphabet

The link here is to the alphabet section in the lexicon. The specific page linked here lists off the Paleo Alphabet with links from there to single pages for each letter and punctuation.

Those pages are built using a template, a new trick for us. Each page has the same form, but is populated with letter specific data coming mostly from the audit system.

Within each page is a long collection of letter related data that matters to people learning the alphabet.

The drawn form of each letter and the Qu Map location start off the page.

Then each page lists off ALL of the kerning pairs for all the letters that can follow any given letter. Correct letter kerning of the text is critical to the recovery process. Kerning determines the pace of advance of text across the audit canvas. So incorrect kerning causes false problems for the audit.

The kerning values do not appear to be given anywhere in the 3d system. Kerning pairs are not hard to guess, but each guess needs to be tested against real words.

This is particularly true when letters can overlap at corners, like Ge before Lu. Those tables show all of the current kerning data values in the audit system. This is our place for visually auditing every single kerning pair.

Beware: Unfinished Kerning Calibration

Note that none of the data shown there has been tested yet, everything is subject to revision. This is like a machine in a factory being brought up for production use for the first time.

Every aspect of the system must be calibrated and verified. I have been setting up the software system so those places of control exist in obvious places that are easy for us to adjust as we get started.

Auditing Data

Below the kerning data, those alphabet pages also contain an explanation and test cases for the audit pattern. Each letter has some theoretical set of audit tests that must be met when those letters are found in running text.

Each letter is drawn in each test location against the audit pattern. Audit failures are marked.

This system naturally lives on the Shepherd's staff item from the Tent of Time. The software here is following that item's design very carefully.

Anyone interested in understanding the audit pattern should read through those pages. Note we tested the audit pattern against the 10 commandments last year. We were using a general computer drawing program and plastic models. Our tooling last year was fuzzy. The art on the pages in the lexicon is geometrically perfect. (Within computer screen pixels and SVG drawing issues, 3d modeling is better.) The art on these pages is showing audit failures that we hunched would be so using the rough designs last year. Very nice to see.

The Lexicon is showing how we are turning to software for testing and correcting the entire manuscript. We have a plan of attack for this which I will be detailing in future blogs. Hint: we intend to start in areas heavy in the use of names.

Lex: Root

The Lexicon also has pages for every root word known in the PEI. This is still under construction, so only about 1/2 of the roots are in the Lexicon. None of them have been audited, again beware.

The link here is to the index page for the roots. That page lists the first letters of all words. From each of those letters there is a link to another page that lists all words by that same first letter.

Lex: Root/Yo

The link here is to the secondary page for all the root words that start with Yo. There is some data for each word. The word itself is followed by the same audit marks in the PEI. Then some stats on the various forms for the word. Each row is itself a link to yet another page.

Joshua

For a final example, the link here is to the lexicon page for the name Joshua as currently spelled in extant Bibles.

The format for these pages is something that I would expect to grow with time. But the key data is already there. In particular it shows how the word is drawn using the current kerning data. It then shows the word in all 6 possible positions along the drawing/auditing canvas. For each letter it shows black circles that mark audit failures.

As currently drawn, Joshua is YoShaVeOo, a 4 letter name. When we worked on the commandments early last year we found we needed a 3 letter version of this name, YoShaOo, to complete the commandments. Word corrections like this will come to these pages as we continue with the work.

As currently shown, the word only passes audit 5.

Finding and fixing all words like this is the long hard work of the recovery process. I would guess many months of hard work is ahead. What I am doing now is just showing the basic tooling for that recovery process.

Aramaic Word Structure

The Aramaic and Syriac words that we find in the manuscripts passed to us by history are composed of 3 different parts. The first part is a set of zero or more prefix letters. The second part is a root word. The third part is any of a limited number of suffixes.

There are some words, like Bethlehem, which are compound words which makes this model a little more complex.

In any case, learning to read the language involves learning the root words and then being able to spot those roots from within the larger actual words as written on the page. This is not easy as many roots start with letters that are also commonly used as prefix letters.

The link to the Joshua page above shows that there are various prefixes but no suffixes. This is typical of names. Let me give you another example which has a long list of forms, including many different suffixes

Father

This is the page for the Paleo root word for father, WaBa.

On that page you can see this is a 2 letter root. It currently passes audit at 0345. Remember, kerning is not yet calibrated, and the Wa and Ba are likely to squeeze together by 1 cell. With different kerning, all the other data in the system will adjust.

But look down on this page. You can see that there are 84 different drawn forms of the word as originally found back in the Aramaic. There is a table at the bottom of the page that shows each of those drawn forms. By inspection you can see (using current kerning) that all of them are legal in at least 1 place along the canvas of running text.

Also on that page you can see that beside no prefix, there are 6 possible prefixes on the front of the root word for father. Let me explain a little more about those.

Prefix Words

The prefix letters function in the grammar as single letter, stand alone, words. We suspect, but have not yet proven, that those prefix letters really are stand alone words in the inspired Paleo grammar.

As there are never single letter words in the Aramaic, we can safely parse off the prefix letters without confusion. All single letter words in Paleo are always prefixes on the following multi-letter root. There of course can be more than 1 such single letter prefix word at the front of an Aramaic word.

In the top parse row of the PEI you can see those single letter prefix words. They are broken off from their roots and include a Paleo dot punctuation after the letter. The auditing process will prove, or disprove, our hunch on this point.

Example Prefix Words

Remember, each letter has a definition based on the 3d model behind the letter. Let me give the prefix letters found before "father."

Ba is a tent. What do people do with tents? They go into them. "In" is the usual prefix word meaning when Ba is used as a prefix. The page indicates we know so far that Ba is ahead of father 13 times in the text. The english for that is simply "in father." The suffix provides the rest of the English for the full meaning of the word.

Du is a purse or sack. What do people do with purses? They open them to find out what (or who) is inside. The page indicates that so far we know of 289 places where Du is used ahead of Father. "who father" or "what father" would be the translation to English. (As in an expression like "what father said.")

Ve is a tent peg. What do people do with these? They tie things to the ground and together. The page indicates that so far we have found 68 places where this is used ahead of father. The word "and" is the normal English for Ve. "and father" is the English for this example.

Prefixes combine, and the page indicates the prefix "Ve Ba" are currently found ahead of father 2 times. "and in father" is the English.

The final prefix letter used with father is the Lu. It is a shepherd's staff, often used to point, or to direct sheep. The normal English word is "to". The page indicates "Ve Lu" is known so far to precede father 2 time. "and to father" is the English.

Then Lu alone has been found so far to precede father 239 times. "to father" is the English.

The suffixes on root words are more complex than we know in English, but they pick up words like "the" or "his" or "hers" as well as various plural forms. I am not going to cover them all here, the English rules are more complex.

Let me give a couple to get you started. A Wa, at the end of a word, is usually the English word "the." But in English we want that word ahead of the root. So WaBaWa is "father-the", or in English "the father."

Fe is a masculine possessive, normally "his" in English. Again we want that word ahead of the root. So "WaBaFe" is "father-his" or in English "his father."

The PEI is using a slightly shorter font for the suffix letters. This is against the normal letter drawing form, but it is helpful to new learners. The same notation will come to the LEX here soon.

We could use colors or other marks to indicate suffixes, but we are reserving the current color and marking system of the BRB for eventual decoration in the PEI.

Statistics and Coverage

The manuscript in the PEI is currently covering the OT and NT. It has around 425,000 total running words. There are about 50,000 different unique word forms.

We have parsed about 13,000 of those unique forms and have found about 4,000 unique roots. We can estimate the number of roots to approach 10,000 once all 50,000 unique forms are parsed.

Of the 425,000+ running words in the PEI, over 300,000 of them are currently parsed. Early parsing work focused on the most common Aramaic words, so we are ahead of where we would be based on averages alone. Those early parsing values were not high quality, so there are some troubles with the Lexicon that will clean up as we progress.

Parsing Process

The work of recovery involves various stages. The audit process comes after the words are parsed. We have brought back earlier work on the parsing process. We are not starting that from scratch. Let me explain.

We began work on the parsing of Paleo words back in 2021. At the time we were working on a Paleo Bible app we called the Spice. It was intended as a critical edition. That work required combining the various Aramaic texts we have in our collection into a single reference edition. From there we set out to parse the roots

The fundamental mistake at the time was to think we were qualified to create a critical edition. How were we to know which manuscript was correct? By voting? Remember the lesson of Micah against the 400 prophets of Baal. Voting is a false path.

Everything we did to combine the manuscripts was a guess. Those guesses risked making matters much worse, instead of better. There was no way to track our guesses and undo them later if needed. So we eventually stopped working on the Spice.

In 2021 we did not have the audit pattern. We only learned that pattern in early 2023. With the audit we have a standard for picking between manuscript variations, though it will also reject much of the Bible as uninspired. So we do not need a critical edition of the Paleo text of the Bible. If we are working an audit somewhere, and have trouble, we will run the audit on the other manuscripts in our collection to see if we can find a recoverable variation.

In any case we realized that building a critical edition was a fundamental mistake, especially given what we knew at the time. We put the project into our vault and left it alone for the past 3 years.

At the same time we were working on the Spice project we also worked out a system for finding the parses for most of the words in the text. I wrote a considerable library of clever code for that work. It too, went into the vault.

I remembered that parse system one morning here recently and went to see if I could find the code in the vault. Would the code still work? Not just the Javascript, but the data flow? Could we use it now?

What a find. The parsing side of that work was not that bad. I brought it back from the vault and just picked up the parsing work where we left off 3 years ago. It is plugging into our current work just fine. You can see the parses from that system in row 1 of the PEI. Let me explain a bit of how it works.

English Parallels

We do have 1 copy of an Aramaic Bible which includes an English interlinear, I gave a link to that work in a blog many weeks ago. It neither provides a lexicon nor does it provide the parsing data we need to build a lexicon.

But the English there is very wooden. So by keeping parallel English and sorting all the paleo words in the entire Bible alphabetically, the parallel English usually indicates enough to figure out the parse breaks for prefix, root and suffix.

My own software from 2021 is doing considerable book keeping to make this easy and to check work. That software, with manually entered parse breaks, is were the parsing data is coming from in the PEI.

So that system gets us the lexicon entries as seen in the LEX app. Again, more work to finish it, but the work flow is understood.

The next problem is assigning English to the Paleo roots. For this part of the problem, that English interlinear is not a very good starting point, and it is copyrighted, so we need another answer.

Is there a better way to fill in the English for the Lexicon? Yes. And it would work with other languages, like Spanish.

The Duluth Algorithm

Many years ago we were looking at the need for a Hebrew Lexicon. This is when we thought we would be using the Hebrew OT for recovery. We had no Hebrew lexicon then either, but we did have various English translations from Hebrew.

At the time we took a drive from Washington State to visit friends in Duluth, Minnesota. I distinctly remember where I was driving when I was hit with a possible way to machine generate a Hebrew English lexicon given any OT English text and a copy of the OT in Hebrew.

When we got to Duluth, I took a couple hours and typed in a prototype of the algorithm to do this. In those days it would be called a multi variable regression analysis, a sub discipline of statistics. (Good thing I once did that sort of stuff at work.) These days it would be called Ai training.

The theoretical matrix size is the number of Paleo roots on one side, and the number of English words on the other, so a matrix of maybe 10,000 by 10,000. Each verse represents fuzzy experimental data populating the matrix. Anyone who knows stats, software, and the Bible, should be able to take it from there.

I had forgotten the specifics of that old code, but that too returned one morning in the past couple weeks after a review of a college text book on Statistics. The English for all of our Paleo root words can be filled in from the BRB manuscript given a reasonably simple bit of probability based software.

The only trick is the quality of the answer is based on how carefully the verses are lined up between the BRB's English versification and the versification used in the PEI.

Ryan was already on that problem before I remembered that Duluth trick. He has been cleaning up versification across all of our manuscripts for use in the VR. When that cleanup is done, we'll use the Duluth trick to populate the Paleo Lexicon with English.

By using the BRB as inbound data to the Duluth algorithm we also get to carry in much of the keyword related tagging found in the BRB. This is particularly important because the set of words that makeup all the names in the text become an important first phase of recovery. Names have a more irregular set of letter pairs and can be used to audit the kerning data, which is also a task still to be done.

The basic tooling for recovery is coming together nicely. The apps out this week are evidence and a hint to what is going on behind the scenes. There is still very much work ahead.

2024-08-15 Watch Date

We have been watching for a possible prophetic headline around August 15, 2024. This would be the story of Moses killing an Egyptian when he was age 40. Moses did this initially in secret, but he was eventually found out.

The 2 working fronts in World War III are in Ukraine and in the Middle East. Barring a surprise from somewhere else, these are where we would expect a matched prophetic headline. Iran attacking Israel was the more expected event for this date. But that front has remained mostly quiet. Iran may be waiting for Jewish Sabbath on Saturday, we will see.

I am currently looking at Russia for the fulfillment of this, likely their military response to remove NATO invaders in the Kursk region. That invasion begin around August 6, 2024, in a lightly defended area along their 1000 km front with Ukraine.

Ground invasions into Russian territory are rare, Hitler and Napoleon are the best examples from history. Curiously, this Kursk invasion involved German equipment with Nazis driving. History repeats. So should the response.

Some have suggested this may have been a Russian trap. It is known to have drawn Ukrainian forces away from the Donbas which is more important to the Russians.

Both the Hitler and Napoleon era ground invasions were answered in the same way. So we should expect the same here. First, the Russians evacuate all civilians. This is still ongoing and taking time. Remember, Russians are never in a hurry.

Once civilians are out of the way, the Russians then go scorched earth, destroying everything foreign still inside their territory. This strategy matches the prophetic story we are looking for, burying their enemy in the sand, like Moses.

The Russians do not report in real time what they are doing on the battle field. So it may take another few days to see better reports of what happened at this prophetic date.

Headline Review

What follows are links to articles that caught my attention this week.

Starliner Saga

The link here is to Scott Manly giving an update on Boeing's Starliner, which is effectively stranding astronauts at the ISS. Their 8 day mission is looking more like an 8 month mission. Boeing continues to look incompetent, and NASA looks like a bunch of politicians.

Organizations tend to face the same problems every generation. Biblically on average generations are 19 years. The 1986 challenger disaster, 38 years ago, or 2 generations back, was NASA management ignoring engineers who told them that the ship could not fly with sub-zero outside temps. Those engineers were correct, frozen o-rings did not contain the rocket's fluids.

The 2003 Challenger disaster was NASA management not addressing the known problem of loose heat shield tiles. They only listened to the engineers that time after the crash, not before. This was 17 years after the Columbia disaster.

This time around, 2024, the problem is NASA management asking an overpaid defense contractor and space buffoon to build a competitor to SpaceX Dragon. At least in this generation, NASA management is not just sending their stranded crew back without thinking about the problem. This is a small step forward for this generation of NASA leaders. But, their crew is not safely home yet.

Boeing wants to fix their problems by going to law against their subcontractors. Boeing is not sure what is wrong. Boeing will take over a month to restore automated control software used earlier on an uncrewed flight. (Insane.)

Depending on how bad this gets, the defective Starliner could damage or destroy the ISS itself. Finally, it looks like Boeing does not have a rocket to use for launching Starliner on future missions. What great planners. It does not look like they ever wanted to fly this rocket. They only wanted the NASA cost plus contract.

There is nothing kind to say about any of this. Let me ask the obvious, was NASA management ever competent enough to reach the moon? Another question, the real question, Is the USA itself loosing Joshua's favor?

The 100 Year Bluff

Very many people have been studying the Covid vaccines since about 2021. Effectively, an army of citizen scientists are at work on this problem.

Turns out that not only are the Covid vaccines deadly, like we did not know that already, but all normal vaccines are dangerous too, especially when used together.

The link above contains a video clip by one of our favorite commentators, Catherine Austin Fitts. The X post also explains the situation along with various links. All vaccines are themselves part of a 100 year old medical bluff.

Nobody should touch them, least of all small children, who are usually damaged for life after being vaccinated. Autism is the most common obvious ailment, which does not exist in the children of non-vaccinated communities. It gets worse. There is no scientific evidence that vaccines were ever safe, as an employee of Gates himself eventually discovered.

Use of Ivermectin

That same wave of study about how to treat Covid vaccine injuries is turning up new uses of the drugs banned during Covid. One of them is the use of high dose Ivermectin to fight the rapid cancers caused by the vaccine. The link given here explains. (I am no expert, but it looks like Ivermectin has a healing mode about like Vitamin C.)

O'Looney Turning Away Business

Our favorite whistle blower on the medications used early in Covid was John O'Looney. He runs a funeral home. Part of his job is to pick up dead bodies from hospitals. In that job, in 2020, he had access to the off-the-record observations of medical staff who were killing their patients with overdoses of a specific medication. Those overdosed medications caused deaths that looked like flu.

In the UK, in 2020, this was ordered done by national staff in London. Real Covid deaths only began later when the Covid vaccine was made widely available. Those same common but very dangerous medications were probably used in the USA to kick off the pandemic here in 2020. There really was something in the punch bowl.

Now O'Looney is reporting that this summer, 2024, his funeral home is overwhelmed with the deaths of young people. In normal times, young people do not die at all and only a few elderly people die in the summer. But if they are dead by vaccine, then there is no longer seasonality to death. The link above explains.

Bangladesh

The former president of Bangladesh has come out publicly and stated that the reason for the coup in her country was because she refused to cede sovereignty over Saint Martin Island and refused to let America hold sway over the Bay of Bengal.

Zionism wants to conquer the whole world and rule it from Jerusalem. She is just 1 more victim. Someone stirred up crowds against Joshua and against Paul. Those same someones still do it today, even in Bangladesh.

Kursk Oblast Incursion

The link here is to the Wikipedia article on the August 6, 2024, incursion by Ukraine into the Kursk Oblast. This article gives the day by day details over the following week or so. This incursion is being blamed for a serious change in Russian thinking over their war with the collective west. Russia is not going to be playing nice any more. Russian leaders are now publicly saying the west risks incineration by their actions. This is perhaps just what Pharaoh wants.

Becoming Full War

This link is to Pepe Escobar's most recent article about the escalating wars. Near the bottom he quotes Russia's Medvedev on his change in stance after the Kursk incursion. Times are changing.

More Later,

Phil