I make lots of small changes around the site. Some to the design, content, maybe a new page under construction, or to some code you can't see. It's not always worthwhile putting a note on the page that an update was performed, so I'll dump all the updates I make on this page instead.
Newest appears at the top.
Update 16/01/2017: Updated the MD5 database information page with more up to date information regarding the bruteforce ranges. This is correct to the best of my knowledge. The only way it could be wrong is if I forgot to include a particular range; the shown data is definitely correct.
Update 12/01/2017: Updated the MD5 database page with information about updates. Just trying to keep people informed. Another great update is that this site now supports Secure Sockets Layer, or SSL! That means encryption via the HTTPS scheme is now possible.
Update 05/01/2017: Working on updating the MD5 Database API a bit. A number of new extensions available to use.
Update 04/01/2017: Done some minor updates on the MD5 Database documentation. Working on each section bit by bit to bring it up to date. Various works have been happening such as allowing the history database to store the hash's algorithm post-crack to exclude it from further processing but the results won't be available to the public due to this service remaining an MD5-only service. Certain API extensions will indicate if the submitted hash was found to be another algorithm (and will indicate that it was not found purely for ease of implementing the API to your own scripts).
Update 17/12/2016: Two new replacement fans have arrived and have been installed. This caused about an hour or two of downtime. I also cleaned out the radiator fins and other parts from dust build-up.
Update 15/12/2016: Had some downtime from 2:20am to 10am due to thermal shutdown. With a liquid cooler, I have a two-fan radiator where both fans have died. This increased the CPU temperature by about 40°C to 50°C and did eventually thermally shutdown. I have two backup high power TFC fans on the case which I'm using now to force air in the case and out of the radiator. Sorry for the inconvenience.
Update 27/11/2016: These last few days have been spent ironing out some bugs and issues I've found within the inter-process communications side of things regarding the MD5 database's recently found passwords table. I have a list of things I need to improve on so I'm slowly working through those. The loop that checks everything seems to get slower and slower over a few days, as if something is building up per entry and not getting cleared and it's wasting cycles going over whatever data is left over, even though I thoroughly make sure all data is cleaned up after use. Give it time, I'll figure it out.
Update 17/11/2016: The recently found passwords table in the MD5 database beta page now supports visible spaces and even puts in a faint dotted underscore to indicate this. Other forms of whitespaces such as tabs, carriage returns, line feeds, etc. will be treated as a control character and replaced to its equivalent control character 'badge'. Ampersands are also replaced. Unintentionally, UTF-8 is treated as-is and remains displaying as UTF-8 encoded Unicode so all those international alphabets are supposedly supported too! My next task is to get hyperlinks working for contributors in the table.
Update 16/11/2016: It's come to my attention that someone (or some people) have been implementing my free MD5 database API in paid software specific to hash cracking. Just to put the message out there before I update the information page: That is not ok. The use of my database is to be free under all circumstances. If you're developing software of which you want payment for, then don't use my API. You can of course explain to your users that they can try through this site if they wish but I do not appreciate or allow people to ask for any payment from other people if my database is involved.
Update 14/11/2016: Those that use the MD5 database a lot might have noticed anything 4 bytes and under doesn't get found instantly and has been this way for a year and a half (since June of 2015). I'm glad to announce this issue has been fixed. All 1, 2, 3, and 4 byte binary values are now included again. This issue arose a long time ago due to a drive failure with no backup. I didn't care so much since they were just precomputed and can be regenerated again, sadly I was just too lazy to actually do that. Within that year and a half, just over 35,000 hashes where not found but could have been had I regenerated that data.
Update 07/11/2016: Progress is a bit slow again. Just been tweaking bits and pieces related to the inter-process communications and MD5 Database pages.
Update 30/10/2016: The inter-process communication for the server and live updating of the recently found passwords table on the MD5 database page is working beautifully. Just need to work on minor things such as timeouts and character escaping of non-printable results. Updating of the number of viewing clients and submitting clients.
Update 23/10/2016: Been on holiday for a while, now slowly back to work. I decided to create another database for every possible byte of 5 bytes in length. This will result in over 1 trillion entries. Due to the new inter-process communication thing being developed, I should be able to manage multiple databases containing various data simultaneously, and through multiple machines for the future.
Update 06/10/2016: The inter-process communications thing is coming along nicely. Doing lots of error checking and internal clearing up and closures. Working on the protocol and how everything will tie in to the website functionality. It's quite interesting.
Update 01/10/2016: So after some time of thinking away from the computer, I decided I need to step up my game and work on inter-process communications for the recently found passwords list in the MD5 database.
Update 17/08/2016: Added some HTML comments to the MD5 database page urging people not to automate the HTML form as it will soon be disconitnued in favour of the dedicated and public API.
Update 15/08/2016: Well, I may or may not have forgotten to keep on top of the changelog. Nothing interesting to really catch up on. Some minor changes to the MD5 database projects but functionality has not changed. Just minor background tweaking and updating of text. Lots and lots of hashes have been cracked and found. I plan to ditch the HTML result page for MD5 results and will make it dynamic with the already-in-situ API. I urge people to use the API because it's already there! Yet people are still scratching the results from the HTML page. I will put out plenty of warning before I make the changeover. Estimated time for the changeover may be in a month's time. Been very busy with real life lately.
Update 22/04/2016: Long time since I've done anything with the site. Sun is shining, summer's coming, and I'm being more social these days. For now, I've tidied up the MD5 history database as it contained some incorrect or unsanitised values which didn't display on the site or API correctly. Despite not having made any changes on site, I've still always focused on the MD5 project and always cracking hashes.
Update 10/03/2016: Alright, so I still haven't had time to fix the factorial calculator yet, so I'm reinstating it with the same problems it had. At least some use is better than no use.
Update 04/03/2016: Someone had reported a problem concerning the factorial calculator page. Quite strange, too. It relied on the GMP library to compute the results, but causes a failure on certain large numbers (but not all large numbers) and seems to stem from within the function that formats the calculation to a string. Unfortunately I don't have the time to look into it and I have removed it from use until further notice.
Update 28/02/2016: Removed the bug fix notice on the MD5 database. Over the past few weeks, myself and a few helping hands have managed to crack almost 20,000 hashes! It's safe to say the development of the larger MD5 database is on hold until further notice.
Update 24/02/2016: Fixed a bad output in the XML API for the MD5 database, the raw password element outputted the byte NCR instead of the byte itself. I fixed it and now it outputs the raw bytes encapsulated in CDATA tags. The input range for the Collatz conjecture script stated that the user must enter 1 or above when in fact entering 1 returned an error. This was pointed out by a visitor who sent me a message about it. Thank you! I've also been using the day to get the MD5 database form to work using AJAX and the self-provided API, going well so far.
Update 18/02/2016: Added a display to the MD5 database recently found password table to show how many people are viewing the content live as it refreshes. This will soon be accompanied by controls to stop and start the automatic refreshing. I've spent the past few days coming down hard on a lot of the hashes that have been unfound for a while. I've managed to boost the overall completion of all submissions up to 82%! Many cracked hashes have also been found to be non-MD5, such as NTLM, LM, truncated SHA1, 256, and 512. Many are also salted despite numerous mentions on the site saying only regular unsalted MD5 hashes only. I also cleaned up some files in the root directory.
Update 15/02/2016: I've added a column to the MD5 database's recently found passwords table to display the number of hits that hash has received. It updates along with the "when" column.
Update 12/02/2016: The automatic updating of the MD5 database recently found passwords has been working great, however I noticed it would stop updating sometimes due to XML errors caused by a PHP error that occurred due to too many database connections logging the requesting hostname and user agent string. These data are recorded for analytical purposes to see who and what my audience is. Though, to be honest I haven't used the collected data in over a year. I have suspended the collection of this data until further notice, and as a result, the updating table works perfectly.
Update 02/02/2016: Successfully implemented an automatically updating table for the MD5 database's recently found passwords. It displays data just like before (escaping unprintable characters, and escapes XML's CDATA's ending character sequence) so there should be no problems. The refresh rate of it changes reflecting whenever the last found password was, since that pretty much indicates someone's using it and there are results being found. Sometimes nothing gets found for hours, and the refresh rate will drop to a minute interval.
Update 01/02/2016: All the API output formats work. All that's left is to output any error details in some of those formats. For now, they'll return a blank document if any real errors occur. Now, I'm working on the recently found passwords list to make it automatically update using jQuery.
Update 16/01/2016: I've spent pretty much the whole day working on the new centralised MD5 database library (for my own use of course). It's early hours now, so I think it's all done. Just got to implement it in to the API page. The API page is now working as it was before! Good stuff. The internals are completely different but it behaves the same way and is now much easier for me to work with.
Update 15/01/2016: Unfortunately the API script (only the API script thankfully) for the MD5 database got wiped blank as a result of my editor crashing. It's not the first time my editor has crashed, but it is the first time [over 6 years or so now] that it has ever lost my work. I'm using this opportunity to recreate the API script and employ a centralised version of my MD5 database for easier management of code. Whilst I'm doing this, I'll add support for multiple output formats including TXT, XML, YAML, JSON, PHP serialize, CSV, TSV, and so on.
Update 08/01/2016: The server went down for some unplanned maintenance so the site may have been inaccessible for about an hour or something. The server now has upgraded cooling. I have a couple of 252 CFM case fans on the way. Additionally, I have only just noticed a visible error with the Collatz conjecture page. An error from the statistics scripts was showing, but didn't interfere with anything. That must've been there for a while.
Update 04/01/2016: So I've spent the last few days actually cracking some of the MD5 database unfound hashes. I've had some good results so far. My plans for this project will include regular brute force cracks of ranges that I don't have stored. Multiple hash submissions (and for the API). A pattern detection thing to see if a result may be salted or SHA1 or MD5 obfuscated; it'll list possible candidate passwords.
Update 31/12/2015: I've found 20,000+ passwords to hashes that were not previously found. Many of which are double MD5 or SHA1 and MD5. I told you people to submit regular MD5 hashes! I'm thinking to periodically run some brute force processes on the unfound hashes every X amount of time (days, weeks, maybe some months depending on the values chosen).
Update 29/12/2015: I have only just discovered that, because all submitted hashes go into a separate database for tracking statistics, the accompanying plain text password of all numeric values have been altered by SQLite3 and have been storing those numbers incorrectly or changing them somehow. When a user makes a second request, the password is pulled from that database to reduce load on the main database, therefore returning the incorrect numeric value, despite the SQLite3 columns being set to TEXT and not NUMERIC. Ah, what a headache.
Update 28/12/2015: I changed the MD5 database table of recently found passwords; instead of sticking the CSS in each DIV tag, I decided to see how compact I could get the HTML and use some CSS3 selectors to find and apply the appropriate style parameters.
Update 26/12/2015: I updated the site banner ready for the new year celebrations.
Update 25/12/2015: So the past few days have been about organising the new database, so there's always stuff happening. I've put up a notice on the MD5 database page about 4 byte passwords or less being temporarily unavailable due to the catastrophic drive failure back in the beginning of November 2015. The external service at md5db.net should pick most of the printable ones up at least. Additionally, I've made a small table of the 10 most recently found passwords on the MD5 database and I've rearranged some of the information on the MD5 database page so there's less text. Oh, and Merry Christmas everybody!
Update 22/12/2015: That fix is still holding up. I believe I can say it's no longer a problem! Additionally, I've changed the website banner to a more Christmas-y feel. We haven't had a very cold winter in London. Temperatures are still between 10° and 15° Celcius so it's hardly frosty (which was the previous theme which also lacked any real Christmas content).
Update 22/12/2015: It appears the fix is still holding. Brilliant!
Update 20/12/2015: I believe I have managed to fix the server offline issues. It kept dropping offline due to Apache becoming unresponsive every few minutes. It's been up and running for a few hours without any faults now. Now we wait for the test of time.
Update 20/12/2015: I've changed some code around and fixed a couple new errors that have popped up in places. I've also added some more hashes found by mamleader, this guy doesn't give up easily. Big thanks!
Update 15/12/2015: I will be adding some provisions to the MD5 database to allow storage of passwords that contain null bytes (which is used as the separator byte). This allows the database to store pretty much any data I want. It has only the drawback of taking just microseconds longer, and uses just a little more disk space. There won't be many entries that actually contain null bytes so it's not a worry. I know many of the recently added passwords actually do contain null bytes, but hey, it's an extra feature for no real cost. This also means I can store all my 1, 2, 3 and 4 byte databases in one large database. One less thing to worry about with less code.
Update 14/12/2015: Added a good 12,168 found passwords from the MD5 database history. I'm quite impressed by this submission because it finds a significant number of salted hashes as well as double hashes of different algorithms. Very good work there thanks to blandyuk from HashKiller.co.uk. I've managed to find about 8,300 few passwords myself weeding out predictable passwords that follow a pattern which were hashed as SHA1 and then MD5.
Update 13/12/2015: I've made an API for the MD5 database and displayed its usage on the MD5 database page. It checks the history database and updates the statistics fine now. I'm also thinking to allow different file extensions on the end of the API URL to allow the results to be delivered in different formats. Additionally, I'm going to slap together all the MD5 database passwords and other passwords I have laying around and finally do something with them all. I have so many collections just laying dormant on my drives. I'm well aware that my server occasionally stops responding; network is fine, it seems to originate from my Apache module. Restarting Apache seems to get it going again despite localhost working completely fine. I removed an Apache module possibly causing this problem.
Update 12/12/2015: Once again, mamleader has added some more results of MD5 hashes to the database. Thank you!
Update 11/12/2015: Added a few more thousand found passwords to the MD5 database, again, thanks to mamleader. I've also changed the way it stores binary passwords in the history logs. Additionally, I've cleaned up some code relating to searching external sites so it's more modular (no point explaining in detail, nobody can see how it works anyway). Instead of immediately deploying the script to search external sites, I decided to play my cards right and request their permission to request data from their sites. Fingers crossed. Upon their acceptance, I will disclose which sites they are.
Update 09/12/2015: There's always stuff being done! Poking around on the Interwebz shows the site md5db.net to be quite popular, or at least active in the way of generating, storing, and finding MD5 hashes. They quite openly provide an API to use, and I have decided to give it a whirl. Their API page spits out a PHP internal error if you don't pass a user-agent header to it; easily to overcome. My MD5 database scripts now use it as a last resort, and saves any finds to its own database and creditsthem for any found hashes. It's now evening time and I got back from work to find that mamleader (one of the trusted users to help find some hashes) has given me 4,043 extra found passwords and has also made me aware of a script error in the MD5 database page which is a result is very late night coding making the MD5 database check external sites; I clearly did not test it extensively. All is well now. I have submitted all the unfound hashes through to md5db.net and got back some rather complicated passwords, only about 30 of them, but I suspect they were generated (and thus stored) on that site by their users and they've submitted them to me to see if my database could find them.
Update 08/12/2015: Again, more slow progress on the ICO extractor. Instead of reading loads of articles, I decided the best way to figure it out and understand it better is with actual samples so I've been making lots of different types of icons with varying bit depths and dimensions and working out the padding bits. It's a bit of a headache but at least this way I know that whatever files it'll receive it should be able to handle. The ICO extractor now supports 8-bit images of any arbitrary dimension. The code behind this ain't pretty for sure, but if I get it right then I may be lucky enough never to have to look at the source again. I just need to enable transparent pixels and inverted colours (to be replaced with another colour). I have removed some of the old update notices from the MD5 database and factorial calculator. In addition to the MD5 database, two lovely IRC users by the name of Szul and mamleader have helped find some unfound hashes for us. Between the two of them, they have cracked 26,445 hashes.
Update 07/12/2015: Alongside the creation of the ICO extractor scripts, I've started work on an informative page for some of the basic PHP scripts I've developed. From really simple functions to the more advance stuff. As far as the ICO extractor script goes, it fully supports 32-bit icons of any arbitrary dimension.
Update 06/12/2015: I've started working on the old ICO extractor script. A script that reads old Windows *.ico files and extracts all the icon images from within them. They contain multiple images rendered at different dimensions and bit depths to allow better control of their display instead of letting the user's machine attempt to display them and possibly display a badly rendered image. Despite how old the Windows icon file format is, they are still a bit complicated to parse and make readily available for the Internet. A lot of icons users had uploaded in the past did not conform to specifications but were still readable by some applications that allowed tolerances in their values, I'm aiming for this also. As far as I can tell, there's either no compression used in Windows icon files unless it's a PNG image despite the fact these icon files are based on BITMAP-type format that appear to allow compression of various types. This makes things easy, however I will set up a notification system in case the scripts detect possible defects or unexpected values in Windows icon files.
Update 30/11/2015: The factorial calculator is back up. A funny thing with this is that it used to take a long time to produce these large values years ago on older equipment. I got around this problem by storing milestone values on disk and would read them then calculate up to the user's input. This worked very well. Since the server rebuild a year ago, I continued to use the factorial database and all my old scripts. Since I had lost all the data earlier this month I needed to reproduce all the values again. I've been figuring out ways to compress the data better, all kinds of algorithms and timing methods. Several hours pass by before I realise this new server build can actually produce any factorial in a mere fraction of a second anyway. Remarkable. I may not even need a precalculated database for this tool. I most definitely will need one for calculating higher values (which I have been mentioning for a while now). I've also tidied up the backend scripts up a little, and added the information pages for each tool to the main site menu but they will only be visible in the menu if you're on the actual page, with a nice little information icon. Repositioned the star next to the donations page menu item. I've added some extended Latin alphabet letters to the NTLM case corrector. And, last but not least, during the late evening, the server went down for some minor spontaneous maintenance. I cleaned out all the equipment, dusted the fans, grills, filters, organised the storage devices, tidied up the wiring, labelled the HDDs as well as removed the dead HDD that caused some issues at the beginning of the month. Everything is back up and running smoothly and dust-free. I expected the GPU, CPU, and PSU fans and heatsinks to be clogged with dust and dirt but they were all surprisingly clean. If you've read this far, I'll let you on in a secret that I'll be working on a couple of
Update 28/11/2015: I've been busy with real life stuff lately. I noticed all the entries in this changelog for November were stuck on October. I've updated the website banner to celebrate Christmas which will probably come and go very quickly. I hope I can actually work on the factorial calculator soon. I don't like my tools being offline. Unfortunately, I will be doing a lot of overtime at work to catch up on funds so I may not have a lot of time left for the website, but I'll try my best.
Update 23/11/2015: I updated the factorial calculator page with a more up-to-date offline notice. It says I should begin working on rebuilding it starting from the weekend. It may not be up and running until some days later.
Update 15/11/2015: At some point yesterday, the drive I'm generating the brute force passwords to decided it was too full mid-generation. Got to continue on another drive, now.
Update 14/11/2015: Generating yet more brute force ranges. I'm aiming for that sweet trillion mark still. The recompilation will take a good month, might as well make the most of it. Trying to fill in a couple of blanks will take some time too. I'd like to just mention that my thoughts go out to France who have been targeted in a series of attacks around Paris last night while I was asleep.
Update 13/11/2015: Reconstruction of the lost MD5 database data is finally complete! The process has made me more aware that improvements can be made and a fraction of it has been lost to do unknown circumstances, I shall rebuild those in due time, it's such an insigniticant portion.
Update 12/11/2015: Reconstruction of the lost MD5 database data is almost complete. Service availability should be 100% by midnight at a guess. Whilst juggling with a social life and work life, I'll get started on the factorial calculator whenever I can. Unfortunate the database for the factorial calculator turned out to be a complete failure, but it's not as big or as complex as the MD5 database. Then, after that, I shall recompile the MD5 database, a fresh, clean, and significantly larger database. I shall also work on it being expandable so that I don't need to recompile the entire thing each time.
Update 10/11/2015: So the reconstruction of the lost data has been hit and miss over the past couple of days. It's been failing after several hours of repetitive work and I can't find out the reason. I've enabled the MD5 database form since at least most of the data can be accessed by now, though not all of it. My script is reconstructing the lost data again, and parts of the database will be accessible whilst it is reconstructed. If it fails again later on, nothing bad will happen, but I will be able to resume it where it left off. Coincidentally, all the brute force patterns I intended to generate are all complete and will be present in the next database compilation. The current database has a lot of duplicates, too, with some passwords not belonging where they should, I honestly don't know how this has happened, but I will ensure it won't happen next time. Additionally, I've fixed the hash generator page, when you click the presented link from MD5 database, shows your found password in different algorithms. There was also a single error in the Collatz conjecture page regarding statistics. I've simply muted it for now, I've more important things to focus on right now.
Update 09/11/2015: The reconstruction has proved effective, but not quite
over yet. I need to test that nothing is missing, that all the data points to the correct corresponding parts, blah blah
blah. Boring stuff. I'm also generating another 91.4 billion passwords in a brute force range. This is in the pattern of
zzzz99999, and from
Zzzz99999. This should
cost about 850 GiBs of space before compression.
Update 08/11/2015: The script I developed yesterday has managed to reconstruct most of what was lost, but it's also given light to some corrupt/inconsistent parts of the database I didn't know existed. There's no actual "loss" as far as I can see, but there's unexpected data where there should exist meta data for example. This caused the process to halt until I saw what was happening. I'll see what I can do about this in the future, I have too much to work on until I worry about that.
Update 07/11/2015: Unfortunately, it seems like this HDD with some of the server files has decided to be more trouble than it's worth. It could quite very well be an issue with the circuit on the HDD itself due to the nature of the fault. Data on it seems to be unrecoverable at this moment in time. I've recovered some important website files, but not enough to put the tools back online for public use. I've been up all day and night trying to fix this issue and now I have work in the morning. I can only apologise for the inconvenience and keep trying. Some of the data can be reconstructed but we're talking days to figure out exactly what the data is as well as the time it'll take to produce it, plus trying to deal with work and as well as trying to keep my already neglected social life. Have faith that everything will be back up and running soon. At the moment, I'm developing a script to read the whole database in order to reconstruct the lost mapping file.
Update 06/11/2015: I have only just realised the MD5 database is also suffering faults also related to the disk storage problems. It's one of my older and smaller storage disks that contains all my older version website data. I shifted some things on there while I was managing the install of the new 6 TiB drive, but now it's dropped out of service. A strange coincidence, but I need to wait until these MD5 database brute force ranges complete. This is just an unfortunate case of bad timing. Sorry, folks. Additionally, in the YouTube CC downloader page, the list of languages failed to show both the Chinese (Simplified, China) and Chinese (Traditional, Taiwan) flag icons. YouTube changed their language codes, and I have updated the code to continue to display them.
Update 05/11/2015: Well, it seems like maybe the MD5 database hasn't been working for the past couple of days. Quite possibly due to the fact that I deleted some old unused files that were being "checked" in the old scripts used to initialise the database. As a fail-safe, it's stopped itself working as those files no longer exist. I've no idea how I've over-looked this, or why I've even used those old scripts at all when all the new scripts did the real work anyway. All fixed. Additionally, I've cleared up about 60 GiBs from the OS drive, and I've added an extra 6 TiB drive to the family, I have no excuse not to try and fill this baby up with the MD5 database. I've also just renamed the title bar text of this Changelog page. I have also changed some information on the factorial page since I have received another message yesterday with someone motivating me to fix the limit (from 100,000 to 200,000). I have also thought about raising the future factorial goal of 1,000,000 to 10,000,000 using an emailing system. In preparation for breaking in the new HDD, I've decided to produce some more brute force ranges! I'm guessing I'll end up with something like 800 billion words after everything has been put together. I have taken the factorial calculator offline due to some disk storage problems, but I can't do anything about it for a day or so.
Update 02/11/2015: All the generation has finished and I still have disk
space left. I have just enough left for another range! I'm going to generate the range from
Aaa999999. This should weigh in at 352 GiBs, adding an extra 37.9 billion to the count. I foresee another
6 TiB drive coming soon. Additionally, I cleaned out the server so it's now dust-free, filters cleaned, and replaced 2
of the slow chassis fans with 4 high flow ones (rated at 130 ft3/min each). I've considered opening up a
donations page; the cost of equipment comes out of my own hard-earned money after all. After consideration, I've finally
Update 31/10/2015: I've decided to go the extra mile and produce another
brute force range for the MD5 database (along with what is in production since yesterday). I'm now producing all
passwords where first 6 characters are the lower-case Latin alphabet, with 2 digits appended. This ranges from
zzzzzz99, weighing in at an extra 260 GiBs uncompressed, with 30.1 billion
Update 30/10/2015: I've changed the background in the site header and added
some extra text in celebration of Guy Fawkes night; one of my favourite celebrations of the year held on 4th of
November. I've added a line in the MD5 database information page, under the brute force ranges section as I'm currently
generating every numeric password at 11 digits in length. This will use about 1.1 TiBs of disk space prior to
compression, and will add an extra 100 billion passwords assuming that none already pre-exist in the database. The range
produced will be from
Update 29/10/2015: I can't seem to figure out why the MD5 statistics is missing some parts out and giving incorrect statistical values. For the mean time, I whipped up some code to check which parts were missing and update them so they'll always be there (but they shouldn't be missing to begin with).
Update 28/10/2015: Similarly to the update yesterday regarding the MD5 database statistics, I have done the same with the Collatz conjecture page. Their statistics are collected in the same way, and are now available to view.
Update 27/10/2015: I have enabled the display of statistics information on the MD5 database page. The processing of statistics had to conform to the recent MD5 history database. It operates a lot quicker than before. Instead of processing statistics from the whole database, it processes only the parts that had detected changes. Later in the day I discovered this process keeps missing parts, resulting in lower figures. This has been fixed.
Update 26/10/2015: The MD5 database was offline between 02:15am and 03:45am. It should operate a lot quicker now, and should sustain higher performance under heavy load. I've also updated the way it displays plain-text values; it now shows the control character values for non-printable bytes. I've also moved some system files around out of the public domain. It's neater now.
Update 22/10/2015: Since people love to abuse the MD5 database page, I decided I will add in some request limits. I have found out some people are requesting the form over 120 times per second and is causing an overall performance decrease on the site. Commencing the 26th, I will disable the form and improve the MD5 statistics. I've mentioned it before, but I can't do anything while people are busy with it. I've decided to announce a formal advance warning about it going offline on the 26th, then I will begin work. Additionally, I have enabled some old databases to the MD5 database. Every byte combination of 1, 2, 3, and 4 bytes in length are now fully supported. Everything is operating fine. I've updated the factorial calculator page just to say that I had only one visitor contact me to get the limit shifted from 100,000 to 200,000 and that if I get another message about it then I'll actually do it.
Update 19/10/2015: I have temporarily hidden the statistics section on the MD5 database page. There's a couple of people making several requests per second and the statistics fetching caused pages to take a second or two to load. I will soon update the statistics database (as I did with the statistics database on the Collatz conjecture page). Additionally, for the last few days, I've been working on the old hash generator page, most of the time spent on this is going towards adding algorithms, comparing test vectors, porting code, etc., I have released the page, but I have warned of possible wrong outputs for some of the lesser-known hashes that I've tried to implement with no test vectors. At the moment, I have 94 algorithms supported. I've added the update to the main page.
Update 14/10/2015: I added a link to the main index page to this changelog. I don't think it's of any significant importance to add to the main menu on the left of the page. I also fixed an error that appeared on the MD5 database page when searching for a new unfound hash. It appears I'm getting a bit sloppy. Additionally, I've been working on the old Hash Generator scripts. I should be able to get an incomplete page out soon.
Update 12/10/2015: Fixed some errors that appeared on the MD5 database page whenever a user submitted a new hash. This happened as a result of an update I made on the MD5 database week that displayed who found the password to an older hash. I also fixed some errors that appeared on the Collatz conjecture page with regard to the statistics of the submitted starting integer. I queried the database incorrectly and yielded no results. I also handled result-less variables incorrectly too. Rookie mistakes.
Update 10/10/2015: I've been working on a new page for helping people speak numbers in different languages. I have English supported, but I wish to allow it to have better support before I release it.
Update 09/10/2015: I have split up the Collatz statistics database in to smaller pieces. Processing statistics for individual integers should be a lot quicker. Processing of the whole thing needs to be planned and finished, though it's not a priority at this moment.
Update 01/10/2015: I have changed the style and position of who found hashes, if someone submitted a hash that was later cracked by someone.