Setup Assistant

Hi, my name is Xsss4hell!
I'm a developer from germany and actually know Pascal, C+, Java and most web-languages.

I worked as Co (Network-) OEM Administrator when I was 17 and noticed that there are way too many steps to finish a system installation.
So I decided to code a Setup replacement that easies and boosts the installation process.
Now I'm 20 and know some programming languages (note I'm still a beginner).

I want to create the ultimative "Setup Assistant" for any Windows operating system. (xp in the beginning)
A program that converts a general windows cd's setup to a dynamicaly assisted setup.
A Setup replacement that detects and installs any hardware just like the original setup.
But assists and recommends you options relating to given information in realtime.


What I wanted was something that melts system updates, tweaking, optimizing, automation, shrinking, customizing and all the other cool stuff to one application. The only reason for this was that I saw too many projects that are amazing but twins of other projects, I think it would be cooler, if the devs would work together instead of coding two or more apps that do the same.. In example: All the "Unattended Makers".


The endusers can be home-, power-, oem- and corportate users
Before you start asking when is it ready.. there's no planned finishing date.
I'm sure this will take long time to become available.
And I hope there'll be some good c++ coders, who want to support and help me.
The reason why I'm askin here is that your project called "WinBuilder" is different in some way to all other projects I know (in a good way) and I like that, but it has also some similarity to my project I'm planning.

Features:
main:
- Setup Assistant
Collects neccesary installation details and automatically, recommends best options.
It's complexity varies from 1 to 10 customization steps.
It'll offer local and remote installations services.
The assistant will help you customizing your installation or just automate and simplify it.
This is done before the setup has begun.

system:

- hardware detection
- general setup routines
- open setup api
- servicepack integration
- hotfix integration
- driver integration / management (driverpack)
- addon integration (updatepack,firefox etc.)
- components management
- tweak management
- xml presets management (import, export / upload, download)
- gn8, il8 (globalization, internationalization)
- network access (tcp/ip,udp,p2p)
- proxy support
- db access (sql,mysql)
- enhanced security options
- unlimited scalability (modular interface / addons, plugins)
- unattended switch finder assistant
- skinable
- customizable
- quick preview window


+This project will be opensource at "sf.net", "berlios.de" or "code.google.com" when milestone 1. is reached


General project timeline:
- finding a name
- finishing planning phase
- creating core functions
- creating a gui
- integrating functions into the gui
- extending features
- adding plugins
- developer feedback
- optimizing
- design tweaking
- creating a project installer
- finishing project
- creating homepage
- user feedback
...other features coming..

In general can you please add any information you know about the windows setup?
In example I know very few people who know what setup.exe really does (and who're keeping it a secret).
I acutally don't know a name for the project except:

--------------------------------
SetupX-10
^X-10 sound just good...
Setup .Live
^(instead .Net
--------------------------------


PS: I hate people crying "that's impossible".
I know much about project-planning and frameworks and so..but don't know what you're actually using.
I'm sure that u would never have any problems with unicode in example, if you would use standard frameworks and libraries. like ICU from IBM...and in general can you tell me please what coding language you'd prefer for coding this? I assume not JAVA, am I wrong? however Java is nice..C++ is a bit more complicated in my opinion.
I tried to plan that project with sereby, but he could not really help, because he knows VB only.
Now he's planning something similar in VB..I think you know him, he's the creator of the german updatepack.

Your project seems very interesting, winbuilder is coded on delphi 7 - mostly to ensure backward compatibility with older OS's.

If you need some webspace on boot-land.net, feel free to host your files around here as well..

Good luck with your development!

I my opinion, your project is ineffective, I mean efforts vs result It's also, not "exotic" enough to be really interesting You're going to kinda ease humiliating installations necessary MS to get paid It's honorable intension, but why not to make "portable XP" that would run on any PC and forget about installations? That would be real job for a man
Seriously, I believe "portable XP" can be done!
I'm sorry if I hurt your feelings, just wanted to be clear and straight
To my excuse I put many to soften it

Alexei

Well,
maybe Alexei, notwistanding the 's, did exceed a bit in putting you down.

@Alexei, you could add a few 's to further make your post diplomatic.

But unfortunately the core of it is right, your idea, while admirable, is actually a bit ambitious, you are basically saying that you want to re-write from scratch the least documented part of the Microsoft OS, the SETUP routine.

I really don't want to be considered one of those

people crying "that's impossible"

but I have to warn you that though NOT impossible, it does appear very, very difficult, I would put it down to a list of priorities, you are perfectly right when you say that a lot of time, knowledge and talent is often wasted in "two or more apps that do the same", but IMHO, your project creates the risk of subtracting more time and talent from other projects that have higher priorities.

Finally, and please do not take the following as a negative comment to your "General Project Timeline", it is only a consideration coming from experience, but you risk falling in one of the hundreds of projects you can find on sourceforge.net that remained in the planning stage or pre-alpha, you missed, IMHO the basic two points:
FIND ENOUGH willing and knowledgeable people willing to SERIOUSLY help you
UNDERSTAND what setup.exe really does and document it
once you have succeeded in these two, EVERYTHING else will be a breeze

Yeah, sure these are the reasons for your post, but they are also the non-avoidable prerequisites.

Unforunately, I know next of nothing about SETUP, so I cannot really help you in this, but if you need any help in MBR, partitions, boot code, drive geometries, fields in which I have some knowledge, feel free to ask anytime.

jaclaz

Thanks for the replies

Nice ideas with that "portable xp" thing, but there's already a project doing that (you could add bashrat's driverpack to it to make it work on any system)

It's called XPE look on sf.net. And yeah I like that projects idea, but look at the project's page and you'll see why I wanted to create this project..XPE project consists of too many subprojects, it "looks" as if the've no plan. I like to have everything ordered from the beginning, before things start getting complicated.

hihihi I know that sereby already has replaced original setup.exe and it works , but there are still problems with hardware detection..but nothing that is hard to do..just to answer your question it not impossible.

setup.exe accesses an api...just replace setup.exe and c what happens ;-)
ok cu l8ter it's late..I'm tired..

Hi there Xsss4hell!

Sounds like you've got a good idea. Customizing XP installs is still a black art. Would be good to have another tool for the common man.

But, before you get too far, what would your program do that n-Lite doesn't?

http://www.nliteos.com/

Take a look

Thanks for the replies

Nice ideas with that "portable xp" thing, but there's already a project doing that (you could add bashrat's driverpack to it to make it work on any system)

It's called XPE look on sf.net. And yeah I like that projects idea, but look at the project's page and you'll see why I wanted to create this project..XPE project consists of too many subprojects, it "looks" as if the've no plan. I like to have everything ordered from the beginning, before things start getting complicated.

hihihi I know that sereby already has replaced original setup.exe and it works , but there are still problems with hardware detection..but nothing that is hard to do..just to answer your question it not impossible.

setup.exe accesses an api...just replace setup.exe and c what happens ;-)
ok cu l8ter it's late..I'm tired..

@Xsss4hell
I assume you meant Sherpya's XPE http://sourceforge.net/projects/winpe
and Bâshrat the Sneaky who wrote this http://www.uawiki.or...d...d&s=bashrat
I have the same impression of "XPE" (and mess with BartPE plugins in general).
Also, real XP that can run on any PC is kind of statement
Ironically, technical base for it is rather close to what you want to create
You can treat startup of "portable XP" on a new hardware as a super-fast setup without following re-boot (of course, next time on the same PC it would boot as "normal").
"Portable startup" can be based on the driver that would run when hardware is already detected and info written into the registry, but before Win begins loading "specific" drivers. At this point your driver can update the registry to make following stuff running smoothly. That's serious job, but it seems to be less then getting full control over setup process The gap between XPE and nLite has to be filled
BTW, is your project supposed to be (and stay) freeware? (Sorry, if I missed you mentioning that.)

Alexei

@Moon Goon
NLite actually does some features I've planned like optimizing, tweaking,updating etc., but it's not a setup replacement and not opensource at all.

@Alexei
Yes I meant Sherpia's XPE.
I'd exactly the same idea with that driver slipstreaming!
Did you know that I don't need to code that if I had PE 2.0?
Vista PE already has an inbuilt function called driver slipstreaming.
I mean devices that were detected during setup are dynamically slipstreamed to a setup package.
I'm planning to do something silmilar.

M$ has thought about the same problems as we do.
The proecess would speed up if the setup would load drivers during boot dynamically into ram and save it directyl in flat registry files, instead of importing every singly entry. What else can we speed up?
Harddisks are faster than CDs...hmm copy the whole setup into ram and copy setup files onto hd. then start setup and decompress neccessary files? I think i need more details..have u any ideas?

My project was intended to help computer noobs to install "xp" with everything set up. The only thing he/she should have to do was to enter his details and select some properties. Then I heard from a friends problem
with installing many computers unattended. He needed an alternative to RIS.

Thats why I wanted to help ie. Network Admins to install one computer's preset to hundres/thousands of other computers. I know some one who'd installed 30.000 computers in a goverment project at once! He'd to code something similar to my project (PXE Boot, Setup Assistant/Replacement, but forms were filled automatically by RIS). He worked with RIS in BIG style..imagine if he'd my tool he would be able to go further then a default installation and application install (RIS has limits..) and on the other hand he would save network bandwith and speed the process enourmously up, because every computer would know what to do instead of waiting for commands.
Compressed files are faster to copy then non-compressed files.

My project will be open-source .
I want others to improve and help that project and life costs already enough, doesn't it?


That' why I've to code "some" features nlite already has from scratch. (you should know that..yea the 1st post was a long one



--off topic
Some devs told me that nuhi doesn't want to share his sourcecode, and he ain't really cooperative if you plan a project with him..but that's just rumors.

I'm not sure about winPE 2.0 since I haven't tried it out yet, but I'm quite sure that copying the CD using a PE environment can be much faster..

Why not running setup.exe from this boot CD on PE mode?

This would also allow to have SATA drivers, detect network hardware and such..

Hmmm..., cannot see how it would be possible to include WinPE 2.0, which is closed-source and non-redistributable in an Open Source project?

And I am perplexed too by the possible "target customer base" of the app, I don't think that there are that many Admins that need to deploy at once more than a few copies of OS, and there are already tools for this either Freeware or Commercial.

jaclaz

I still didn't get it
What range of scenarios you want to cover with your software?
I can hardly imagine numerous ITs installing OS via P2P
Hardware were much weaker when MS was designing its installation process, why follow their path?
For ex. It may be simpler to exclude excessive drivers when Win already detected the hardware Though, why really bother?
If process is fully automated, it's not that much important how fast it runs. They can do it overnight What really matters is how much manual work is involved, so it may be interesting to make completely automated setup initiated by network boot
You may be interested to look at "Slimming Down Windows XP" (though the guy is kinda strange)
http://www.bold-fort...hp?showforum=13

Alexei

@jaclaz
I didn't want to use PE 2.0. I just wanted to explain that and how others solved that problem with slipstreaming drivers from setup to hd.

@Alexei
My project was planned to boost the user through the setup process, while giving the possibility to customize his/her final os before installation has started. Later I thought about Network Admins problems.


So, it seems as if some people don't believe that this is possible or what I say..?

OK..

I HAVE already an application that saves XML presets and can do some options. It runs on PE autostartup and collects information, then gives it to the setup. Rest is unattended.
But I wanted to go a NEW and BETTER way. TELL me any alternative to Ghost, RIS and Sysprep to make custom installs unattended..none or nothing comparable, especially nothing FREE?? hmm..that's why I wanted create something into this direction..by removing any unser interaction during setup and doing any interaction on beginning of setup..the rest should work automatic.
Now don't say NLite...or can your NLited image work on an other computer. I mean the computer has an other user who wants an other user account and apps, settings etc..

Now there are left two alternatives classic way: MS Setup, or this new way...

I told you that I know somebody who already finished a project similar to mine.. here are more details:

To clear things up: I don't like the RIS method.. I want a new method which is more flexible and not causing so much traffic etc..

********************
He'd to install about 30000 XP-Cliens nationwide. (don't ask he's not allowd to give it to 3rd party)
He's working for a big security agency int the IT-secor in germany.
His project is based on a RIS-Server per Location. The clients boot a WindowsPE image over RIS.
But instead of viewing the default WinPE-Shell interface a GUI appear written in C appear wich does following:

- Collecting information computer-name, domain etc.
- How to format drives?
- Which Sofware shall be isntalled?
- Which system components shall be installed?

The GUI creates an unattended.txt and opens the setup's unattended routine.
He'd to write some tools to join a Domain and connect Network-Drives
It works like a charm

And WinPE already detects a big amount of Hardware (Networkcard drivers are most important for RIS), so that's definately the better way of installing over RIS.
********************




grr

Would a person be able to use this app with PicoXP? Sounds good from looking at the pictures.

@Alexei
My project was planned to boost the user through the setup process, while giving the possibility to customize his/her final os before installation has started. Later I thought about Network Admins problems.
So, it seems as if some people don't believe that this is possible or what I say..?

You know, I still have strange feeling that you don't have a clear picture of what you want to create
Is it gonna be similar to those pictures?
If yes - that's very good, but this guy had specific task with limited variations of environment, conditions, software, etc. So, he could make some "standard" setup and let users to mark what they want. You can't go the same way unless you define your target
What would be typical usage of your software? What variations of typical usage you're going to support? What kind of users you expect? What you don't want to support and why? I still don't know your answers

Alexei

Unattended RIS installs:
http://www.msfn.org/...p?showforum=127

Unattended project
http://unattended.sourceforge.net/

RyanVM's integrated update packs
http://www.ryanvm.net/msfn/

Bashrat the Sneaky's DriverPacks
http://www.driverpacks.net/

N-Lite
http://www.nliteos.com/

XPize
http://www.msfn.org/...p?showforum=112

I think what you're talking about is doing something that replaces all of the above. The thing is, RogueSpear is doing that right now. It's called AutoImage:
http://www.doitright...e0a87e6a2a83b87

The screenshot you're showing is from Client Installation Wizard for RIS server for Windows: (You can customize it)
http://www.microsoft...b.mspx?mfr=true