Neptune build 5111.6

[Lukas Marsik]

In late 2015, the disc of Neptune build 5111.6 was spotted in a video published on Microsoft's Channel9, called Defrag Tools #143 - Raymond Chen - Microsoft Archives Part 1
The disc makes a short appearance around the 15:45 timestamp as people in the video tour the archives.
Here's a crop of the disc in question:


Earlier this year I happened to get in touch with an ex-Microsoft employee that worked on Neptune, Millennium and Whistler.
He said that he most likely still owns the last build they got to take home, and after a few weeks of waiting a CD labeled Neptune 5111.6 12-99 indeed turned up.

Receiving the image was both cool but also slightly underwhelming. The OS files happened to align perfectly with the long leaked "5111.1." Turns out we owned .6 all along.
The .6 you see on the disc's label does not refer to the build number itself, but rather the variant. Microsoft produced a multitude of these with a different set of internal components, registry hives geared for different audiences (such as OEMs), and other similarly small alterations.
In my opinion the variant claim is backed by the existence of a commented entry for "NTC Selfhost" in the list of Startup programs in the leaked copy. This image also reaffirms the legitimacy of existing Neptune images, bar a few missing folders. Subsequently confirming that Neptune media was built by taking a chosen stable Windows 2000 build and overlaying updated Neptune components on top. The base build for 5111 is build 2128. Feel free to check this yourself by diffing the aforementioned build's images -- a lot of files will match.

Another point I'd like to touch on is 5111.6 being the final build. The infamous wAxWaReZ NFO actually alludes to Neptune already being a thing of the past by the time it leaked: He NUKES Neptune !!

Code: Select all

 
 
   ╔═══════════════════════════════════════════════════════════════════╗
   ║                                                                   ║
   ║                   Windows Neptune build 5111                      ║
   ║                                                                   ║
   ╠═══════════════════════════════════════════════════════════════════╣
   ║                                                                   ║
   ║       Format.....:.bin/.cue      File Count....:14 @ 20MB         ║
   ║       Supplier...:Unca Bill      Cracker.......:who knows         ║
   ║       Packaged...:wAxWaReZ       Util Type.....:O/S               ║
   ║       Date.......:12/??/99       Protection....:None !            ║
   ║                                                                   ║
   ║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║
   ╔═══════════════════════════════════════════════════════════════════╗
   ║                                                                   ║
   ║    What does Bill Gates do one week after he puts himself in      ║
   ║    charge of all software development and makes Steve Balmer      ║
   ║    CEO ?  He NUKES Neptune !!  Oh m'gawd !!  What a BAD move !    ║
   ║    I've been using this Developers Release since I got my hands   ║
   ║    on it back in December and it's wonderful ! OK ... so it has   ║
   ║    one quirky bug in explorer that won't open pages in the        ║
   ║    correct frame of a framed web page.  Big deal.                 ║
   ║    The original setupreg.hiv file has been replaced so this thing ║
   ║    NEVER expires !!  Originally it was a 444 day trial that reset ║
   ║    itself when you reinstalled.                                   ║
   ║    Only system need to worry about is make sure you have enough   ║
   ║    RAM.  I started off with 64MB of SDRAM and quickly found out   ║
   ║    I needed more.  96MB is bare minimum and 128MB is preferred.   ║
   ║    If I had a laptop, this OS would be on it faster than you      ║
   ║    could say "BOO !".                                             ║
   ║                                                                   ║
   ║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║
   ╔═══════════════════════════════════════════════════════════════════╗
   ║                                                                   ║
   ║  Installation Notes :                                             ║
   ║  Unrar and burn to CD with yer favorite burning util.             ║
   ║  Doesn't ask for a CD key and as previously noted, doesn't have   ║
   ║  any time limit on it any more either.                            ║
   ║                                                                   ║
   ║  A definite keeper for collectors ---                             ║
   ║                                                                   ║
   ║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║║
 

Even though the image isn't all that different as I've said, I figured I should still share it with the community. You get to see the elusive "NEPTUNE M2" label, as well as a few more 2128 folders that didn't make it into warez releases, presumably due to them being fairly useless and just inflating the size. Original timestamps also make their appearance, for you timestamp fanatics.
Big thanks to the ex-MSFT guy for staying in touch and offering to upload a piece of history. I'll probably disappoint you a tiny bit by saying that he imaged it as an ISO and not superior MDF, but I figured I should be thankful for what I got. There's not much hidden data one could pull out of a CD like that, plus he's a pretty busy guy so I didn't want to drag stuff on.

Here's the link to the ISO: https://drive.google.com/open?id=1FIrup ... vGUtDPnHox_
Probably the best 5111.6 copy we'll ever get to see.

There may be more images of different builds of different OSes on the horizon, but it's not 100% certain for now.

[Overdoze]

To support the "Neptune is an add-on for Windows 2000 RC builds" theory some of you may find hard to believe, here's a news snippet from ActiveWin from August 24th, 1999:

Neptune News
Time: 15:07 EDT/20:07 GMT Source: E-mail Posted By: Alex H

We are trying to get some more screenshots of Microsoft Neptune to post on the site for you to see. At the moment Neptune is in the very early draft stages. In fact, it "lives with" Win 2000 Professional now -- kind of an add-on to Win 2000 PRO (Build 2072 ONLY). Music Center, in the screenshot we posted on the 22nd doesn't even work yet - it's broken. Neptune's version is 5.50 (for now, just like Millennium is v4.90 now).

The screenshot they're referring appears to be this and should be familiar:

[mrpijey]

Very nice find, thank you! This may finally settle a few arguments we've had in the past.... I've replaced the copy on the FTP with this one.

[Lukas Marsik]

Thanks for processing it quickly

[Lucas Brooks]

Thanks for sharing the original Neptune 5111!

Looks like I did get the timestamps of files right but didn't get timestamps of directories right. The original disc images was created using CDIMAGE 2.39z which was unexpected. The CDIMAGE parameter used to build the original disc image was very basic, it doesn't even have an AutoCRC embedded and the optimize storage option was not used.

How to rebuild the ISO to produce identical result:

Code: Select all

1. Have all files ready
2. Change the timestamp of the directory with all Neptune files to 1999-11-17 09:04:28 AM
3. Download my fake CDIMAGE 2.39z
4. Change your time zone to UTC-8:00
5. Change your BIOS date to 1999-12-10
6. Build the ISO with the following parameters:
    CDIMAGE -l"NEPTUNE M2" -n -bBootSect.bin sourceroot targetfile

Download CDIMAGE 2.39 (fake): https://mega.nz/#!gRN1hQCZ!yF406YHHT45G ... A6QRf7txKM

[andrian1994]

Whoa nice find! , Thank you for sharing

[FateForWindows]

Awesome! I'm glad we have a clean copy of this. Hope we get more (earlier) builds someday, though all things considered it doesn't seem too likely.

[anonymous74]

Still amazing, same or not.

[JimOlive]

Thank goodness it's in .iso format. Hopefully that person leaks older builds too.

[YourAverageJoe]

I find it curious on how the label, "NEPTUNE M2" tipped off from the guy who apparently owned a copy of 5111.1 (which if being honest, is probable bull from the terminology 'Software Expiration' and the fact subtracting it from the 444-day timebomb won't produce an accurate enough date) makes it here onto the actual. I feel like further checks should be done to ensure someone just didn't take a Win2000 RC2 ISO and pasted in anything that wasn't from September 9th, since it was known for a while at least from what I read 5111.1 was distrusted through a network share. Don't want you to go nagging this individual any further, but a more cleaner image of the disc would at least prove the legitimacy. Doesn't matter if done through a scanner or a phone camera, just an image that includes the bottom text.

Something to mention about the original scene leaker of what was inspired for the copy on the FTP and the ‘more-original’ take, Pirates With Attitudes, was that they took completeness very seriously. Just read this portion of the NFO to their release of Win98SE, which can be found here.

While every other group will be bringing you so many good programs for this operating system, it's PWA that brings you the OS itself. It is fortunately for the user community that this is the case or you would probably have ended up with a ripped down release from some other lame group missing important system files like KRNL386.exe, because disklimits are more important nowadays to these people than a working release.

People that still believe in quality have only one option : Pirates with Attitudes.

Only reason I am mentioning this, is because the amount of files in I386 folder on the PWA copy matches the amount on the supposed ‘5111.6’ ISO. Yet that’s the only folder it contains, doesn’t contain the others since which why would it? It doesn’t need them, as they never bothered making anything exclusive to Neptune in its place, as documentation in those folders and the interactive discover thing still is implied for Windows 2000.

[andrian1994]

I installed today and find no different , same like build 5111.1
I'm curious what the big different of this build anyways?

[Lucas Brooks]

I feel like further checks should be done to ensure someone just didn't take a Win2000 RC2 ISO and pasted in anything that wasn't from September 9th, since it was known for a while at least from what I read 5111.1 was distrusted through a network share.

This copy is 100% real and original, from what I can see, it is a home-made copy of the original Channel9 disc. I don't think there is going to be any difference if a cleaner image was taken or a better dump was made because the disc itself is not a Microsoft disc and the dump would be a dump of a copy of a disc which means all the sub-channel data and other metadata were probably lost already. Neptune was originally an add-on for Windows 2000 RC1 and by the time of 5111, it was based on Windows 2000 RC2. There is nothing wrong with Windows 2000 RC2 files being there since it was Neptune copied onto a directory with Windows 2000 RC2 files extracted from the official RC2 ISO. I never believed Neptune 5111 came from a network share but I am not denying the possibility and I think it is illogical for an official developers release to be distributed through network shares.

I checked the ISO with a hex editor and it was made on the 10th of December 1999 while the root directory was created on the 17th of November 1999. People 99% of the times neglects the timestamp of the CD root since it is not visible so I doubt anyone would want to spend time faking that.

I installed today and find no different , same like build 5111.1
I'm curious what the big different of this build anyways?

Because there is no difference and that is why PWA took out those useless Windows 2000 RC2 bits.

[Overdoze]

In 99% of Windows NT releases, warez groups (including PWA), only kept the i386 folder (or Win95/Win9x/Win98/Winme in case of Windows 9x), even when the rest of the CD contents could've been useful to someone. Space saving was simply more important at the time and anything non-essential was ripped. They did this with other software too, stuff like clipart, FMVs, etc. was often removed. Sometimes they weren't careful enough and removed something important as well, and then some other group had the opportunity to make a better release and brag about it (which is what your PWA quote is referring to IMO).

So I don't think this proves anything regarding the authenticity of this copy.

[YourAverageJoe]

This copy is 100% real and original, from what I can see, it is a home-made copy of the original Channel9 disc. I don't think there is going to be any difference if a cleaner image was taken or a better dump was made because the disc itself is not a Microsoft disc and the dump would be a dump of a copy of a disc which means all the sub-channel data and other metadata were probably lost already.

Right, would make for a curious tale though. You know, on why a ex-Microsoft employee would do that. To protect the contents of the disc before they get split into two? But I digress

Neptune was originally an add-on for Windows 2000 RC1 and by the time of 5111, it was based on Windows 2000 RC2. There is nothing wrong with Windows 2000 RC2 files being there since it was Neptune copied onto a directory with Windows 2000 RC2 files extracted from the official RC2 ISO. I never believed Neptune 5111 came from a network share but I am not denying the possibility and I think it is illogical for an official developers release to be distributed through network shares.

Yeah, this I had known already. I meant it towards unofficial means, like if this guy could just grab a 2128 ISO and shove anything not from the date 9/9/1999 back inside. The network share is a little curious, sure, but that’s where the discs come in. There could have been both different illations of the same build, one for the testers to make fine improvements that is on the network share and the other (5111.6) to developers when they felt done. Like, when Longhorn has two of the same builds built at different times with subtle differences. And when a release gets fixed under the same build number days after it hits RTM, like with NT 4.0.

I checked the ISO with a hex editor and it was made on the 10th of December 1999 while the root directory was created on the 17th of November 1999. People 99% of the times neglects the timestamp of the CD root since it is not visible so I doubt anyone would want to spend time faking that.

Yes, maybe. The date does match with the supposed compile date on BetaWiki, and the filedate on the directories. The November 17th date does seem odd, but so does the time on the I386 directory of the ISO given to us here. But again, it could be that the files were applied within that date and in turn it changed which explains why the folder dates inside are from like 5 hours and 45 minutes before. It’s also something to note, COMPDATA (the first alphabetically) is a minute before the rest. Showing to me a sign of it being say, pasted onto a network share for people to copy and paste. That’s all this image looks to me, it looks legitimate but I’m positive it was likely mislabeled. Especially considering the odd CDIMAGE version used, was that used in Win2000 RC2 (2128)? If no, then likely supports what I’m saying; network share.

So, too long didn’t read.. Disc may be legitimate, but I very well do doubt it was a copy of the disc from Channel9 with it being on a blank CD more than it was a Win2000 RC2 (2128) disc slipstreamed with Neptune files from a potential network share. It would not make sense for Microsoft to keep on making small differences dependent on OEM that they need a small decimal on the end, and especially is odd that all the files here are the same exact as the releases we all have seen minus the modifications done behind the scenes. I think that 5111.6 was indeed a more spread-out build, while 5111.1 was simply put it, for testers to crank out any last remaining bugs from probable use with the operating system beforehand which is why the network share. 5111.6 on the other hand was an actual developer release and was distributed in close circles, not as widely as other developer releases, and mostly all copies were likely snapped in two after being done. The surviving batch, being kept in close arms like seen with Channel9. This is my theory, as OP’s theory is very odd on how there’s apparently multiple purposes that can be achieved by swapped files (like the registry hives), yet the main focus in a developer release should be developing software and drivers that works well with the platform for when it gets released. So, this gives us the probable idea that this disc is mislabeled and that 5111.6 is actually still out there.

[Lukas Marsik]

Disc may be legitimate, but I very well do doubt it was a copy of the disc from Channel9 with it being on a blank CD more than it was a Win2000 RC2 (2128) disc slipstreamed with Neptune files from a potential network share.

Tell me one good reason to maintain two variants of media that install the same thing

[winnt32]

Right, would make for a curious tale though. You know, on why a ex-Microsoft employee would do that. To protect the contents of the disc before they get split into two? But I digress

Uh, this is just a disc he took home when Neptune was cancelled.

So, too long didn’t read.. Disc may be legitimate, but I very well do doubt it was a copy of the disc from Channel9 with it being on a blank CD more than it was a Win2000 RC2 (2128) disc slipstreamed with Neptune files from a potential network share. It would not make sense for Microsoft to keep on making small differences dependent on OEM that they need a small decimal on the end, and especially is odd that all the files here are the same exact as the releases we all have seen minus the modifications done behind the scenes. I think that 5111.6 was indeed a more spread-out build, while 5111.1 was simply put it, for testers to crank out any last remaining bugs from probable use with the operating system beforehand which is why the network share. 5111.6 on the other hand was an actual developer release and was distributed in close circles, not as widely as other developer releases, and mostly all copies were likely snapped in two after being done. The surviving batch, being kept in close arms like seen with Channel9. This is my theory, as OP’s theory is very odd on how there’s apparently multiple purposes that can be achieved by swapped files (like the registry hives), yet the main focus in a developer release should be developing software and drivers that works well with the platform for when it gets released. So, this gives us the probable idea that this disc is mislabeled and that 5111.6 is actually still out there.

You are clearly not aware of how Microsoft worked at this time. At this time, anyone enrolled in Microsoft's beta program got weekly builds, so that they could develop with the new features. It was also confirmed by someone - who, unlike you, actually worked at Microsoft on Neptune (not to mention ME, Windows 2000 for Alpha, and XP - in the space of two years no less!), and who actually worked 90 hours a week and slept in work during XP - that the revisions were just hive variants. And before you use "memory fading" as an excuse, he labelled this disc in January of 2000 at the end of the Neptune project. Not only that, having hive file differences makes sense - employees may want to use the build for a different time than OEMs, for example.

Please do not speculate on things until you are actually aware of what you are speculating on.

[andrian1994]

Disc may be legitimate, but I very well do doubt it was a copy of the disc from Channel9 with it being on a blank CD more than it was a Win2000 RC2 (2128) disc slipstreamed with Neptune files from a potential network share.

Tell me one good reason to maintain two variants of media that install the same thing

That's what I worried about [emoji849]

[Lucas Brooks]

I checked the ISO with a hex editor and it was made on the 10th of December 1999 while the root directory was created on the 17th of November 1999. People 99% of the times neglects the timestamp of the CD root since it is not visible so I doubt anyone would want to spend time faking that.

Yes, maybe. The date does match with the supposed compile date on BetaWiki, and the filedate on the directories. The November 17th date does seem odd, but so does the time on the I386 directory of the ISO given to us here. But again, it could be that the files were applied within that date and in turn it changed which explains why the folder dates inside are from like 5 hours and 45 minutes before. It’s also something to note, COMPDATA (the first alphabetically) is a minute before the rest. Showing to me a sign of it being say, pasted onto a network share for people to copy and paste. That’s all this image looks to me, it looks legitimate but I’m positive it was likely mislabeled. Especially considering the odd CDIMAGE version used, was that used in Win2000 RC2 (2128)? If no, then likely supports what I’m saying; network share.

CDIMAGE used in RC2 is 2.27 and that is why I was surprised when 2.39z was used. Below is what happened from my analysis of the ISO:

1. Contents were copied from Windows 2000 RC2 ISO to a directory on the 17th of November, copying finished at 09:04:20 AM

2. Everything from the I386 directory were deleted before 10:30 AM 10th of December

3. Some Windows 2000 RC2 files were copied into I386 at roughly 10:30 AM 10th of December

4. Neptune files were copied into I386 some time after 12:43 PM and copying finished at roughly 04:15:20 PM

5. The ISO was built with "CDIMAGE -l"NEPTUNE M2" -n -bBootSect.bin sourceroot targetfile" after 04:15:20 PM 10th of December and before 11th of December

I don't see anything wrong with that as everything makes sense and all process happened in chronological order. 17th of November means they started to prepare the release to developers in mid-November, copying from the Windows 2000 RC2 finished at 09:04:28 AM and only the I386 directory got modified ever since.

That does not tell me it came from a network share because Microsoft did release official discs with inconsistent timestamps. Microsoft just didn't use the -t parameter either because they're too lazy or they are making it easier to separate Neptune files from Windows 2000 files or some other copies (variants) are add-ons without Windows 2000 files or they just didn't bother.

If it was downloaded from a network share, the root directory's timestamp should be after the compilation date but in this ISO's case, it was almost a month before.

[YourAverageJoe]

CDIMAGE used in RC2 is 2.27 and that is why I was surprised when 2.39z was used. Below is what happened from my analysis of the ISO:

1. Contents were copied from Windows 2000 RC2 ISO to a directory on the 17th of November, copying finished at 09:04:20 AM

2. Everything from the I386 directory were deleted before 10:30 AM 10th of December

3. Some Windows 2000 RC2 files were copied into I386 at roughly 10:30 AM 10th of December

4. Neptune files were copied into I386 some time after 12:43 PM and copying finished at roughly 04:15:20 PM

5. The ISO was built with "CDIMAGE -l"NEPTUNE M2" -n -bBootSect.bin sourceroot targetfile" after 04:15:20 PM 10th of December and before 11th of December

I don't see anything wrong with that as everything makes sense and all process happened in chronological order. 17th of November means they started to prepare the release to developers in mid-November, copying from the Windows 2000 RC2 finished at 09:04:28 AM and only the I386 directory got modified ever since.

That does not tell me it came from a network share because Microsoft did release official discs with inconsistent timestamps. Microsoft just didn't use the -t parameter either because they're too lazy or they are making it easier to separate Neptune files from Windows 2000 files or some other copies (variants) are add-ons without Windows 2000 files or they just didn't bother.

If it was downloaded from a network share, the root directory's timestamp should be after the compilation date but in this ISO's case, it was almost a month before.

I find this point the most valid, thank you. Didn’t take too much thought into the exact process and how the dates may differ if it was through a network share, compared to if it was not. I was thinking Neptune, from what I read a while ago, was purposely like that so testers could just copy the I386 folder from there and ‘patch’ 2128 themselves. I should probably stop speculating however, as that one guy said; as wherever I got this kind of information off from was likely also very wrong and I’m just spreading it.
Also, to the person asking for the point of two releases.. Here I ask, what’s the point of the multiple Longhorn builds with the same build number but different compilation date? I do feel slightly defeated in this idea to combat it any further as the points are right, but I’m at least glad for the progress this community has gone through. I was probably picking at it to begin with, because each time we got new Neptune it was in the form of fakes

[Lucas Brooks]

Here I ask, what’s the point of the multiple Longhorn builds with the same build number but different compilation date?

Because they are supposed to stay internal. Those official releases are all from the same lab and same date. It could be a re-compile of a build with the same build number but those earlier recompiles are not supposed to reach the general pubic. Even if we get 2 recompiles of a build from the same lab, the executables and timestamps should be different. If there are different SKUs or copies, at the very least the registry hives should be different. If you look at my rebuilt 5111 and the original ISO, the timestamps of files matches perfectly and all files are identical.