43 replies to this topic

[FesterJester]

I am looking to create a minimalistic Windows 7 PE environment to recreate a previous project I did.
I am looking to remake this project http://reboot.pro/15425/ using a Windows 7 base so I can incorporate all the LAN and chipset drivers that Windows XP does not support.
Also so I can incorporate a secure hard drive erase program.
I notice that xiaopang has made MicroPE x64, but I need to have it x86 as not all the machines I might use this on are capable of booting x64.
I have also built the Win7PE_SE project, but that has more stuff in it than I would like and takes a bit longer to get to the desktop.
I have even had the thought of making the environment from Windows Server 2008 Core as it is already pretty stripped down..I think.
Though I am not really sure how to start making a PE from scratch using Winbuilder.

So, I call upon the awesome reboot.pro community to try helping.
Any help is appreciated.

Thanks,
FesterJester

[pscEx]

Try minimum configured multiPE.
 multipe_barebone.gif   29.4KB   59 downloads
Peter

[Brito]

A competition was announced via newsletter today. The goal is to achieve a cool barebone Win7PE, if you haven't yet received the newsletter on your email box then wait a few hours (at max one day).

Let the games begin!

[pscEx]

I decided not to take part in the competition.

Requirements:
- Overall size bellow 100Mb (compression is allowed)
- Windows Explorer must be available
- Use Winbuilder to automate all building steps

WinRe.wim, which is a win7PE with the only sence to restore the system from CD, has about 150 MB.
No explorer, no extras, no ...

My smallest win7pe is about 200 MB. No chance to fullfill the requirement.

Peter

[Brito]

The goal is to have fun and give the opportunity for everyone to participate.

Already build myself several Win7PE well bellow the 100Mb mark doing nothing else other than starting from a minimal build, removing files and testing in the emulator the result. If it boots, remove some more files..

To help others, I released some time ago the scripts that I use on my own tests: http://reboot.pro/fi...85-bootdisksdk/ This should help others to quickly create new ISO images and test them on a virtual machine.

To speed up the WIM creation, either write a batch script using the WAIK or the reWIM tool to avoid installing drivers on the local host: http://reboot.pro/fi...file/189-rewim/

Good luck to everyone participating!

[Wonko the Sane]

Possibly a stripped down WinPE 3.0 could be a good base (or however be a source of inspiration).
See:
http://www.msfn.org/...0-minimal-waik/


Wonko

[pscEx]

As far as I understand, your link connects to a stripped down WAIK.

Peter

[Wonko the Sane]

As far as I understand, your link connects to a stripped down WAIK.

Well, as far as I understand it, it connects to a way to build a stripped down PE 3.0 from WAIK.


Wonko

[pscEx]

Maybe my age lets me become stupid, but I do not see any hint about a stripped down PE here:
 mimwaik.gif   13.41KB   21 downloads
Maybe in any of the replies there is a PE mentioned. Who knows ...
I usually do not scan a complete topic (actually something about WAIK), to perhaps find an OFFTOPIC post (actually something about PE).

Peter

[Wonko the Sane]

Maybe my age lets me become stupid, but I do not see any hint about a stripped down PE here:
 mimwaik.gif   13.41KB   21 downloads
Maybe in any of the replies there is a PE mentioned. Who knows ...
I usually do not scan a complete topic (actually something about WAIK), to perhaps find an OFFTOPIC post (actually something about PE).

Peter

You see, a pre-requisite for understanding is reading (as opposed to "skimming quickly").
Like if everything posted on reboot-pro, including many projects were aptly titled/described and if you were excelling in being accurate when posting....


Wonko

[MedEvil]

I don't understand the point of creating a PE, which is simply small, without any regard for funktionality.

Wouldn't it make more sense to ask for a PE with at least some real wordl application, like a PE which has the most functionality and can still run on machines with just 512MB RAM?

[Brito]

I don't understand the point of creating a PE, which is simply small, without any regard for funktionality.

Wouldn't it make more sense to ask for a PE with at least some real wordl application, like a PE which has the most functionality and can still run on machines with just 512MB RAM?

It is a challenge. You have years ahead (and before) of you to create complete boot disks. On this competition, it is a category of light-weight boot disks.

In the same way as you don't need a nuclear-powered aircraft carrier ship to cross the Nile river from one border to the other, small boot disks are not full-fledged boot disks and serve well other purposes that you might not understand but that others find useful.

On this category you get points for:
- Boot time
- Overall size
- Features included (drivers, apps, functionality)
- Style (theme, look&feel)

[Wonko the Sane]

As a matter of fact it would make sense to have a PE with the BARE-BARE minimum and a builder/tool that allows the final common user to add easily one by one or by groups only the specific features/tools/apps he/she specifically wants.

I seem to remember there was something like that - at least in theory - that in actual practice became an unmanageable mish-mash of confusing and interlaced pre-sets that invariably result in something almost, but not quite, completely unlike the working build you would expect (when it creates a successful, booting build, of course).....


Wonko

[MedEvil]

Te problem with your small and easy extendable idea is, that a PE is only (more or less) easily extendable by someone, who has a lot of background knowledge, about the inner workings of Windows.

Most people can't even clearly state all the features they need, because they have no clue about dependencies.

Or the user wants access to options A, M and Y. If i don't offer them, the user complains, but if i offer him access to options A to Z he feels overwhelmed and complains that the project is unusable.

So your Winbuilder project would first need an artificial brain, best with telepathic abilities to conter peoples inability to express themselfs properly.
Do you have some lying around?

[Brito]

Yes, complicate if you wish but I prefer the basics.

Just copy files from the source, build the ISO and run on QEMU.

[Wonko the Sane]

Most people can't even clearly state all the features they need, because they have no clue about dependencies.

Well the actual "job" of the developers of such an hypothetic app should first thing have clear themselves the dependencies (or use the Embedded version database as someone suggested in the past):
http://reboot.pro/2690/#entry18030
hey wait, no it cannot be used:
http://reboot.pro/3465/#entry24645
http://reboot.pro/14..._50#entry128510

Also I seem to remember that someone wasn't that happy about the evolution of a tool:
http://reboot.pro/11419/

The attempt to revert one's (or a group of people's) own inadequacy to the "final" user because he/she is a moron by definition - while this latter statement may apply on a case by case basis - seems to me besides improper, also leading to nowhere.

What is the idea?
Educate the users instead of making the tool easier (within limits)?


Wonko

[ericgl]

As a matter of fact it would make sense to have a PE with the BARE-BARE minimum and a builder/tool that allows the final common user to add easily one by one or by groups only the specific features/tools/apps he/she specifically wants.

Wonko,
I believe it's called Windows Embedded Standard.
If you don't check any item during installation, you get a bare minimum Windows 7/8.
cheers!

[MedEvil]

The problem is that easy and flexible don't go hand in hand, but oppose eaxhother.

I can give you a one button project, which is easy to build by anyone.
Or i can give you a highly flexible project, which you can adapt to your every need, but that would require the user to know, what he's doing. As i can't give you a telepathic project.

And as far as, scripts handling dependencies, but not having dependencies, goes. This just can't be done with simple scripts. Unless each script contains every last bit it depends on, down to the bootmanager.

Such a project would take days to compile.

If your idea should have any chance of success. You would need to abandon the script thing and go full fledged programming language, so you can have access to the tools you need.
Also you can't have people just scribbeling their own scripts or plugins.
You will need to set standards, enforce them by thoroughly testing / reading the plugins for compliance and not allowing distribution without your OK.

Aside from the time factor, standards and enforcing, doesn't sound like anything we ever had here.

[Wonko the Sane]

Wonko,
I believe it's called Windows Embedded Standard.
If you don't check any item during installation, you get a bare minimum Windows 7/8.
cheers!

Hmmm, the same (BTW wrong) suggestion, only with a few years of delay.

@Medevil
Comeon, the only difference between a scripting language and a "full fledged programming language" is only that this latter is compiled and thus it will be faster.

And there is nothing "wrong" in making "self-standing" packages (yes, having all their dependencies explicited) if you want in a very little scale (and within the very narrow scopes/target), this is exactly the approach used in the XPCLI and in Misty's MiniXP.
If you prefer there is no reason why the "components" approach so unfriendly implemented in Windows Embedded cannot be made easier to manage and more "granular".

But also someone managed to make a BartPE in 35 Mbytes, if I recall correctly.... :
http://www.911cd.net...showtopic=12067
http://www.911cd.net...showtopic=20608
though the very base approach was - while at the time one of the easiest one - the exact opposite of my idea of a modular PE (or whatever), the approach was based on "subtraction" where I would likeone based on "addition".

If you think about it, a normal NT based system installation has a number of basic functions (and of subsystems), obviusly increased at any release, to which you add by installing them third party programs.

There is in theory nothing that prevents anyone from - once agreed on a BARE-BARE build - to attempt installing on it any original MS subsystem or program or third party program and list the dependencies (all of them) and the needed files/settings to be considered a pre-requisite for installing that particular app or subsystem, nor to have the "install" script verify the existence/conformity of these pre-requisite and if needed create them.

Also, I have seen that a lot of accent has been put lately in "speed of build" (which is a good thing in itself) that I find one of the least important things for a final user, at least personally I would rather have a builder/project/whatever that after havng provided a clear set of steps chunks away doing it's business allright in such a way that when I come back next morning I have a surely working build (but as often happens, it's probably just me)



Wonko

[MedEvil]

For once, i would really love to see your idea of a base 35MB build, which is easily and simply extendable, with all the settings, to a full blow PE, as a proof of concept from you.

I think, you probably have replaced real world engineering with wishful thinking in a few spots.

For second, for most users an easy to use program means, an easy to use GUI.
With a static GUI like in scripts and WB, one has to know, how things work because the GUI does not offer any clues.
A dynamic GUI on the other hand shows, what effect each choice has on the other options or even the program itself.
The best that can be done with scripts in this regard, is a wizard and wizards are highly annoying, when the program is used often.

[Wonko the Sane]

For once, i would really love to see your idea of a base 35MB build, which is easily and simply extendable, with all the settings, to a full blow PE, as a proof of concept from you.

I think, you probably have replaced real world engineering with wishful thinking in a few spots.

Most probably, yes, I am not a programmer like you are , and so it is very possible that what I see as possible it is not, though in my past experience, every single time a programmer (or more loosely an IT professional) said to me "It's not possible" in due time it became evident that the real meaning was either "I am too lazy to work on that." or "I have no time to do that" or "I am not capable of doing that." or "I, being a programmer, hold all the truths and noone that is not a programmer should ever attempt to hint that my view on this topic may be wrong".

But it's OK, if you are happy with the current situation, only issue is that you seem like not .

Anyway, I will gladly accept your word for it , so it can be declared that no further improvement of any kind can be made, we already have the best possible tool and apart producing more often than not builds that are not working or not bootable or flawed and it's use being impossible without a manual:

Without reading a handbook and i really mean BOOK, most if not all of our projects are not to be mastered.

(your very words) we have the best possible program, the best approach, all the best .scripts and any attempt to better anything in them is - like resistance - futile .

For second, for most users an easy to use program means, an easy to use GUI.
With a static GUI like in scripts and WB, one has to know, how things work because the GUI does not offer any clues.
A dynamic GUI on the other hand shows, what effect each choice has on the other options or even the program itself.
The best that can be done with scripts in this regard, is a wizard and wizards are highly annoying, when the program is used often.

Which brings us back nicely to where it started:
http://reboot.pro/in...showtopic=11419
(please do appreciate the new old older syntax I used for the above link )


Wonko

[MedEvil]

But it's OK, if you are happy with the current situation, only issue is that you seem like not

Well since we programmers are all stupid or at least lazy, how about showing your superiority, with your proof of concept?

So far all i see, is a really nice wishlist.

And no, i'm not happy with the status quo, but there's a difference between technical possible but rejected for political reasons and technical not possible.

But as said above, prove me wrong. And don't be afraid to hurt my feelings.
I love to be proven wrong. Opens up a whole new universe!

[FesterJester]

Ok, this thread has gone in a bit of a crazy direction.
@pscEx: I will try a minimum configured multiPE and see if that is close to what I am looking for. Also cause it has really been the only response to my question so far.

How does one start building a WinPE from scratch or am I getting over my head with that?

FJ

Edit: I did update my site to give credit where it belongs. If I have missed anyone in giving credit, please let me know.

[FesterJester]

@pscEx: I am trying to build the minimum multiPE, but am having trouble.
First the ImageIndexEx script doesn't seem to mount the WIM properly, so I have to click "Mount WIM" before the play button.
Second I get "Run - Failed to find section [WimCaptEx_Capture] in file: [%BaseDir%ProjectsmultiPECore1 - Core Basics01-ImageIndex.script]" and the project quits to the log window.
I have verified the script is in the proper folder and that there is no section with that name in that script. I used WinBuilder 082 to download the project files and noticed WinBuilder updated its self to 083 during that time. Am i missing something or is this just a minor mistake that made it through? Let me know if you want log files.

Thanks,
FJ

[pscEx]

Sorry, in the upload 01-ImageIndex.script is corrupted.

Please use WinBuilder's download tab to redownload the script.

Peter