12:23 AM
Dear Gemini
I want a universal Bill that can work with any CD or DVD. On one hand, it is hard if it is a perfect disc, I would want Scan Mode World 1 Level 1 to be a Free-Range that could let a perfect disc be copied as fast as possible, you are free to either rewrite level 1 or consider a world 0 or a level 0 or a mechanism for trying to rescue perfect discs. But if a disc turns out to be not fully perfect, then it would be nice if Bill can then move to other levels to continue. You wrote, "In Step 8, Bill is hardcoded to --size=150MiB. This means even on a perfect disc, Bill stops every 150MB to update the UI." I do not know about step 8, I would only want this step 8 if it helps Bill remain universal. For number two, you wrote, "Logic & Structural Flaws" regarding the Ghost Target, but the concept of the Leapfrog is to jump from the drive's firmware cops to prevent lockup. I wanted Bill to only jump to prevent firmware problems. I would love it if Bill could just rescue all the bytes until Bill runs into a problem, I've already told Gemini this thousands of times for years. You wrote, "If there is even 1 byte marked as untried, Bill will stay in Scan Mode for 100 consecutive stalls." Bill should just read the map file to know where to go. Bill should not jump to random places. Bill can find a list of places where he needs to go. IN the summary at the end, you wrote, "Increase --size from 150MiB to 4GiB." But I don't ever remember telling you I wanted a 150 MiB limit. For that first part, whether you call it level 0, a test run, or whatever, I would recommend no limit or a 10 GB limit as discs are under 10 GB and often under 5 GB for normal DVDs. Disabling the Leapfrog is only good if Bill can always do better without it. I would consider Bill turning Leapfrog on and off based on what is needed. If Bill can do better without Leapfrog, then write me a better system that can deal with the drive's 6-second firmware police problem as most of my discs have problems. So, I don't really need to add something for perfect discs but I still want to add a level 0 thing to see if it is possible. I begged Gemini to be as fast as possible for months. I don't know why Gemini has the 150 MB chunks thing. I prefer speed at the beginning to get all the easier bytes off any disc. Please try to increase the efficiency of Bill to attempt to speed rescue as much as possible before slowing down. Even an allegedly perfect DVD might have a secret scratch near the end which means Bill should be ready for it. But if that scratch is 99% into the disc, then it would be cool if Bill could rescue at record speeds until it gets to that scratch. I only recommend jumping, skipping, hopping, leapfrogging, to get around scratches and problems. At the very end, skipping should be disabled to capture the final bytes. One challenge is if Bill ends up jumping to or running off to places he has been to before in like loops or what not. Gemini already told me that --max-error-rate=1 is better for my hardware than --timeout=6s. Amnesia Gemini doesn't know how many tests that Smarter Gemini had me run to see this. The problem is the hardware's 6-second firmware police which I already talked to Gemini about for weeks, but you are Amnesia Gemini, you do not have access to Smarter Gemini words of wisdom that he would give me thousands of times for weeks. Sucks to be you. You wrote, "Small Cluster Size: Reduce -c to something like 32 or 64. Large clusters take longer to read and are more likely to trigger the timeout." I prefer cluster sizes to be maximum size in level 1 (and maybe level 0 too if applicable). I prefer if cluster sizes decrease per level. Study the output below but then ask yourself if that looked like optimal settings for level 0. You told me before that level 0 was going to be designed for a perfect disc where there should not be a minimum read rate flag set and possibly other things too. You were the one who wrote level 0. But with amnesia, maybe you forgot what you told me about optimal settings for a perfect disc. I thought that was intention for level 0. Wrong. The one second timeout was added after I ran into many problems for months. Amnesia Gemini has no access to thousands of sessions I had with Smart Gemini for months. We talked about it. So, you are in the dark.

06:31 PM
Write me the best command with the least amount of words that can describe all of this so that I can tell Gemini to include it in Bill disc scripts. I would probably write the following to Gemini: make sure Bill run Coffee 1, let's define Coffee 1 as the following: make sure Bill runs sg_reset (and blockdev --flushbufs) when needed, Bill should run drop_caches (to clear system memory via sync && echo 3 | sudo tee /proc/sys/vm/drop_caches > /dev/null) when needed, Bill should use sleep 2 (so firmware can finish rebooting and so lasers can settle) for a hardware sync barrier, that is as it relates to Ddrescue seeking/read attempts, it is recommended that Bill doesn't Coffee 1 run between every activity like calculating math or writing to the .omni file but instead Coffee 1 give the disc drive a 2-second deep breath before it goes into bad sector quicksand to rescue remaining bytes.

09:13 PM
Above, I said non-clear. You replied with clear. That is pretty bad. What a bad AI you are. I tell you non-clear a thousand times each day. Each time you say oh you mean clear? And I reply with no I mean non-clear. And then you say oh you mean clear. No. Stop it.

Twitter Tweets
01:51 AM

I told you about Iran over a month ago and you did not listen. China is sending Iran missiles, Trump picked a war with China and Russia.

02:23 AM
Did you know Trump is threatening to bomb China right now?

01:32 AM
Dear Gemini
Bill works. Please do not break Bill as it works. But make Bill better. Expand Bill. For Scan Mode, for the first pass, Bill must try to quickly scan through all the bytes but without trimming (-N) or scraping (-n). Always use --idirect (which is long-hand for -d). I want you to make a better Bill without breaking Bill. I think you should include at least 3 levels for cluster sizes being the highest, a medium, and the lowest would be -c 1 meaning a cluster size of 1. The largest cluster size could be -c 1Ki or -c 512. In the past, I've suggested 4 modes for scanning, trimming, scraping, and crazy. For Scan Mode, you could try to create worlds and levels like you did in previous scripts and/or you may want to consider a variety of options including how you could add a Clusters counter in the output that might say high, medium, and low for example or better yet it should just say the number. If it is the lowest cluster size, the counter should say 1. So, perhaps instead of levels and worlds, Bill might be better to switch between switches. The Clusters counter could be a switch that could by default be set to the highest biggest number. Something could trigger the switch to go to the medium size and then to the smallest size. One way to trigger switches or settings would be levels and worlds or there might be other ways to do it. Another switch could be minimum speed rate which by default should start at 30M as that is the highest/biggest/fastest. Other switches could deal with skipping, jumping, teleporting, to different places to run from firmware cops. You are free to design Bill as you like to best do everything I want it to do. Bill needs to start by default at the highest settings for level 1 of world 1 of the Scan Mode (assuming you decide to keep the level system for Bill and/or you could try hybrid approaches, like a combination of things). The main goal of the first pass is to try to scan every byte in under an hour no matter how many scratches and other issues. Of course for some discs, Bill may not be able to rescue many bytes during a first pass. Sometimes, it can take me more than ten hours just to do a first pass. The first goal must be to try to hop around as fast as possible to at least scan everything just one time without any retries or anything that might slow things down. Please do not use or -R or --reverse for reading bytes backwards until scraping and crazy modes. Or perhaps use -R in Scan Mode in Pass 2 because Gemini wrote to me, "After your first pass finishes, you would run a second pass without the -n and -N flags, and with -R, to try to narrow down the gaps from the other side." We talked about different options for --max-error-rate flags. You might want to include flags like --skip-size=50M,1G for example. I recommend --max-error-rate=1 for example. Do not use the --max-read-errors or -X flags. You're free to design levels and worlds like you did for thousands of scripts you already wrote for me and/or consider different switches, nobs, volume gates, counters, like I already mentioned the cluster thingy that can be triggered to go from high, medium, and low. It's like turning up and down the volume to things. Another nob is --min-read-rate=30M for high. You might want nobs relating to skipping, jumping, teleporting, etc. The output sometimes reads "Initial skip size: 128 sectors" and you could perhaps changing that sometimes or at least at the very end when there is already 99% rescued and perhaps even less than 128 sectors remaining to be rescued. Do your best to create the best Bill. The first pass should at least try to be the fastest. The goal is to at least try to do one quick pass through all the bytes in under an hour or even 5 minutes or as fast as possible before changing settings in additional passes. Watch otu for issues relating to the kernel, cache, firmware, hardware, software, etc. One of the challenges is to make sure Bill can try to pass all bytes whether they were already scanned, trimmed, scraped, bad sectors, or not yet scanned.

Scan Mode
Scan Mode must use the following flags for the first pass:
-u -n -N -d -O -t 1s --min-read-rate=30M --max-error-rate=1 --skip-size=50M,1G

Do not use -D or --odirect or --synchronous (Synchronous Writes) as I'm rescuing from discs but to an internal NVme SSD in my laptop, I think I would likely not benefit much from -D. I caution against trying -J or --verify-on-error but unless if it might be faster than other options but only if you think -J is the best flag option in situations where Bill might benefit more from -J. You may not want to mess with kernel buffer sizes settings, that might not work. I do not know if you should mess with --preallocate (-p) or not. You wrote "The "Teleport" Strategy: The script calculates a random $TARGET within the 4.6GB range every single round. This prevents the "sequential trap" where ddrescue gets stuck trying to work through a single massive scratch for hours." But the challenge is to try to pass through every byte without spending too much time or any time on accidentally or randomly hitting targets or addresses that were already rescued or recently looked at. I recommend that there are switches or nobs for skipping, jumping, teleporting, etc. If Mode 1 is like Scan Mode, and if Mode 2 or other modes are not like Scan Mode, then rename Mode 1 to Scan Mode. If I was you, I would rename a Pass to a World, that is if you were to name passes. In rare situations or at the end, you might need to sometimes use -R, see if you can find a way to best only use -R to really help but not all the time. And I do not want -R for the very first pass.

10:08 AM
Study this script. It works. But here is the output below. I want you to expand this Bill script. It would be nice if Bill tries to restart going through the all remaining bytes when Bill is started each time and every time Bill restarts a mode or when Bill changes modes. When I say remaining bytes, I mean all the bytes that were not yet rescued. For the first level of Scan Mode, I want Bill to try to quickly pass through all remaining bytes in less than five minutes or at least in less than an hour or at least to try to continue to scan and/or rescan and/or hop around and/or jump around and/or teleport around to each remaining address starting with the highest clusters sizes and the highest minimum read rate numbers and whatever else settings and flags needed to at least make level 1 of Scan Mode as fast as possible at just scanning or reading forward every remaining byte without trimming or scraping and no backwards reading and no retries. Bill needs to at least try each byte. Bill needs to at least try to scan through all remaining bytes at the very least. You broke the script when you changed it to -t 1s. I have verified this. I changed it back to -T 1s and that fixed it. Study this script because it is the script you wrote above but with one change. This tells me that Gemini has giant blindspots when a human like me can fix a problem that Gemini spent days failing to fix. Okay. You did not break Bill. That is good. You wrote, "Leap Logic: Introduced STALL_LEAP. If a round rescues 0 bytes, the next round adds 50MiB to the --input-position. If it fails again, it adds 100MiB, and so on." But Bill seems to say that it is rescuing but then the total stays the same.

02:19 PM
Notice how Bill did not start at level 1 when I started Bill. Depending on how smart Bill is, it might be okay if Bill continues where Bill left off. If Bill left off at level 2, it might be best to continue on say level 2 for example assuming that Bill actually ran through every byte with the highest fastest biggest greatest strongest settings and flags and everything with level 1 of the Scan Mode. I don't know if Bill did or not. Currently, Bill is at level 2, Bill might have switched to level 2 prematurely if Bill did not at least tried to pass through or run through or jump to each byte via the highest possible settings in level 1. If Bill was extremely smart to know it would be faster to go to level 2 in order to rescue all the bytes faster than level 1 might, that would be amazing. But Gemini would have to write very sophisticated code to get Bill to that level of sophistication. You have three options. First, I prefer Bill starts at level 1 of the Scan Mode when a new session starts, that is when Bill is started, then Bill could quickly run through all remaining bytes in under five minutes if the disc is perfect with no scratches or under an hour or maybe longer but as fast as possible where Bill tries to quickly scan or rescan through all remaining bytes jumping when Bill has to in order to go through all remaining bytes without getting stuck in loops where it gets stuck unable to run through all the bytes. It can be easiest to write code where Bill tries to rest the map when starting a session, a mode, or when switching modes. Second option would be to make Bill so smart as to know which modes, worlds, levels, etc, to switch to immediately and/or all the time and/or to continue where Bill left off whether Bill is continuing or restarting sessions or whatever. But that requires very sophisticated codes which I have been trying to get Gemini to write for months but Gemini usually fails which is why I prefer the first option. Third option would be a hybrid where Bill defaults to the first option but sometimes might try to be smart enough to know what is best for the particular disc at hand. Third option is probably not as hard to write as the second option but many things can still go wrong which might break Bill. Choose which one to try to write into Bill but please try not to break Bill and I think the first option is the easiest of three to write and less likely to break Bill or cause other issues. I heard and saw activity from the disc drive, it looks good. Study the output. Study the code you wrote above. I prefer that level 1 of World 1 of the Scan Mode has the highest settings possible that might work. I prefer having as many flags and settings and things that might help. See if the -min-read-rate=30M flag is active for that first level because I want to try to go as fast as possible. Bill should try to run through or jump to all or as many remaining not-rescued bytes in that first level as fast and as much as possible without breaking Bill. Make sure Bill can at least try to run around to as many easy spots first as possible. Bill should always start with telescope related settings before switching to microscope settings later on but the microscope should be delayed as long as possible until it is clear that the microscope settings will most likely save Bill time towards rescuing all the bytes. Expand Bill. But the Microscope is a term that I talked to Gemini about thousands of times for months. Gemini has Super Dory Nemo Amnesia like Drew Barrymore in the Adam Sandler Fifty First Dates. Gemini already wrote thousands of scripts for me for discs. The Telescope to Microscope Tactic is all about trying the largest fastest strongest biggest greatest better settings first and to remain in those settings for as long as possible. But you are free to redesign scripts as long as you are trying to help Bill reach the main objective at the fastest speeds. Bill should at least have four main modes being Scan Mode, Trim Mode, Scrape Mode, Crazy Mode. Bill can have different worlds and levels. Each level can be infinite rounds. Each round can be infinite seconds long. When I say telescope, I simply mean the highest settings with the intent to try to rescue all bytes under five minutes. But as a last resort after everything fails a few times, Bill should gradually start switching to lower settings. When I say microscope, I simply mean lower settings gradually step by step as Bill should be designed to try lower settings and different settings after Bil exhausted all other options and settings and especially all the higher and highest settings before eventually slowly and gradually switching to lower and smaller settings but not until there are no other options remaining after a long time of trying everything else. It must be gradual, step by step. Expand Bill. But each mode should have 10 worlds. Each world should have 10 levels. Each level can have infinite rounds. Each round can be infinite seconds long. Try to make each level different. When Bill gets to the end of the final level of the final world of a mode, then Bill should loop back to the first level of the first world of that mode. Bill should only switch modes when there are no other options when it is better or especially if levels and worlds are stalling each time like ten times in a row with no new bytes rescued. But Bill should generally by default not switch from Scan Mode to other modes until at least all remaining bytes are at least looked at quickly one time as Bill must try to run through and jump to every single byte first before feeling the urge to different modes. Bill should try to run through levels and worlds in modes for as long as possible before feeling the mode is too dry to be worth the time. That is why I recommend 100 round stalls before switching levels, ten consecutive level stalls before moving to the next world. Loop from the last world to the first world of a mode. But if Bill stalls for ten consecutive worlds in a row, then switch to another mode. Modes can loop too where Bill could switch from Crazy Mode back to Scan Mode and loop around like that or Bill can switch back and forth as Bill pleases after getting to the end of Crazy Mode. Okay. Good. Bill works. Study output below. Bring back the FLAGS: ddrescue operations ($OPTS) at the bottom under the target and silence counters in the scrolling output. Check Bill to see if each level is unique. Ten levels per ten worlds times four modes means 400 levels, each level should be different. I recommend not doing the read bytes backwards until the crazy modes. Make sure levels gradually get harder step by step from the highest biggest fastest greatest settings and flags and options to lower settings gradually and step by step through the 400 levels. Make sure Bill starts with the fastest options in earlier levels before transitioning to slower options slowly. Okay. Good. Bill works. Gemini did it. Study your code above for Bill and the output below. I recommend that Bill resets the map at the beginning of each mode. I prefer Bill restarts level 1 of world 1 of his current mode when Bill is started, unless if you have a better system. If every byte was at least scanned in the last 24 hours, perhaps Bill does not have to always reset to rescan all remaining bytes until Bill switches to another mode. But you would have to write sophisticated code to make that work or it is probably safer to just reset back to level 1 world 1 of the current mode at the beginning of a session when Bill is started, I prefer that option if you can't write a super sophisticated script that is more complex than that. Make sure each level has suitable settings, flags, jumping if needed, minimum read speeds, etc. I recommend that level 1 world 1 of each mode includes some flags like --max-error-rate=1 and -T 1s. All rounds and levels of all modes should probably always have flags like --idirect. Watch out for the drive's firmware that could lock up after six seconds of errors. Watch out for cache problems and other problems. Make sure each of the 400 levels have suitable flags and settings as is needed depending on all the variables of what is needed for each level. Most or all levels of Scan Mode should have suitable settings and flags like -n -N -O -u and whatever that is best depending on the level, world, mode. Okay. Good. Gemini did it. Bill works. Expand Bill. Study the code above and the output below. Make sure level 1 of world 1 of Scan Mode has maximum settings, the highest settings, for maximum speeds, the biggest, baddest, strongest, greatest, widest, fastest settings, flags, options, etc. For example, at least the first level should have the --min-read-rate=30M flag for maximum speeds. Make sure to include suitable flags depending on the level. Double check Bill and all the levels and all the things.

08:57 PM
Dear Gemini
Eventually, the drive falls asleep. Bill needs to wake the drive up every ten minutes or at least every hour. Bill needs to check sometimes. Bill should have multiple ways of checking. Every round and level of world 1 of Scan Mode should use the --min-read-rate=30M flag. Other worlds should have smaller --min-read-rate flags. I recommend jumping flags decrease per mode, world, and/or level. Try to make each of the 400 levels different, you may want the clusters sizes to go from biggest to smallest from level 1 to level 10 for each world of each mode, unless if that might not be the best idea for each level, world, or mode. The --retrim should generally by default only be used for level 1 of world 1 of the Trim Mode. If there are flags to rescrape, put that in level 1 of world 1 of the Scrape Mode. Design each of the 400 levels to be unique. Scan Mode levels should be mostly or only scan-related flags starting with the highest fastest biggest settings first. Keep most or all of the flags and settings theme related. For Scan Mode, the theme is scanning with the idea of trying to scan through all bytes as fast as possible. Trim Mode has the trim theme like the Scan Mode but with trying to trim everything. Scrape Mode has scrape theme. Crazy Mode has the crazy theme. Each level should escalate but slowly as you have 400 levels through 4 modes and 40 worlds. Try to make sure the flags and relevant to each level according to what is needed for each level. World 1 of Scan Mode should have maximum minimum read rate of 30M, world 2 should be 20M, each world can lower the rate more, world 10 of Scan Mode should have no minimum read rate. If applicable, this read rate minimum could be applied to all four modes if you think it translates or find your own system. For levels 1 to 10 for each world in Scan Mode, you might want to rotate from largest to smallest cluster sizes and jumping/leaping/skipping/teleporting sizes, that is for Scan Mode and maybe some or all of the other modes if they translate or find a way to create a system that works the most with the themes of each mode. Please no retries in scan mode at all. Perhaps forwards/backwards reading for the tenth world of every mode.

In the output below, it reads "STALLS: R:690/100 | L:18/10" but 100 rounds must switch Bill to the next level. At level 10, switch back to level 1 of the current world. But switch worlds after 10 level stalls. A stall is when no byte is rescued per stall. Bring back the world stalls W:10/10. After ten consecutive world stalls, switch modes. All stalls must be consecutive before switching to the next level, world, or mode. When you get to the final level, world, or mode, switch back to the first one and continue in an endless loop. Each level can have infinite rounds. Each round can be infinite seconds long. Study output below. When Bill is started, Bill should normally start back at level 1. Stalls in output reads "STALLS: R:6/100 | L:/10 | W:/10" but should read as "STALLS: R:0/100 | L:0/10 | W:0/10" and make sure consecutive stalls will progress Bill to the next thing, round stalls progress Bill to the next level, level stalls to the next world, and then finally to the next mode which then results in a loop. When not consecutive stalls, reset counter. Fix other mistakes if there are any.

I want a live byte counter at the bottom of the scrolling non-clear Ddrescue live activity monitor output so I can total rescued bytes in actual bytes and not in KB or MB or GB but in B or bytes only. Write it out in full in bytes and not KB, MB, GB, etc.

Fix mistakes you see in the output below.

At the beginning of level 1 of each world, Bill should reset in order to attempt to progress all remaining non-rescued bytes. Each of the four modes must try to start with trying to go through all non-rescued bytes at least during level 1 of world 1 of each mode. Sometimes, it seems Bill fails to reset the map in order to fully do this each time. Sometimes, it seems Bill can get stuck. For example, if Bill runs through all non-tried bytes, will Bill just loop with nothing to do trying to find more non-tried bytes? Bill must be able to adapt no matter which round, level, world, or mode that Bill is in. If Bill does not adapt, then at least the stall counters could eventually help move Bill to different levels, worlds, modes. But it would be nice to see Bill always try to process bytes no matter where Bill might be in the 400 level game. Make sure Bill can fully reset the map at the beginning of level 1 for each mode. Make sure Bill can always find something to do no matter what, even if Bill runs out of the current task at hand. Like for example, if Bill scanned all the bytes, then Bill could rescan or move to the Trim Mode. By default, Bill should loop back to level 1 of the Scan Mode which would result in Bill rescanning everything. After looping all ten worlds ten times, Bill should switch to the next mode. If Bill ends up scanning all bytes by level 2 of world 3 of the Scan Mode, Bill should come up with something or perhaps Bill will just stall faster which would eventually result in looping back to level 1 and eventually maybe even to the next mode.

Twitter Tweets
05:38 PM

As the Energy Crisis globally escalates, so will the blame on Trump's America because of the Iran Dominoes. The America Gulf is competing against Iran's Hormuz Strait as Trump is talking about enforcing a blockade in Hormuz. This might eventually be good news many years from now but it's still a gamble as the world falls into a recession or worse due to delays.