jaclaz

Are you sure? You need the reg.exe in the 2K support tools (or at least this what has been always used). But you can use REGEDIT.EXE allright manually. jaclaz

No, it doesn't matter on which PC you attempt the recovery, as long as you use in TESTDISK the right geometry. The PHYSICAL geometry of a hard disk (this hard disk particularly) is 255 heads. From the screenshots you posted TESTDISK has it as 240 heads. See: CHS 96901 240 63 So, from what I can say from a distance it is "your "computer that has it "wrong" and on "her" computer geometry is 255 allright. (unless you ALREADY modified data on the disk, if this is the case, cannot say) Let's take first partition as an example. it has 31457280 sectors and starts at 0/32/33, so: (0*240*63)+(32*63)+33-1= 2048 <-start LBA 2048+31457280= 31459328 <- End LBA 31459328/240/63=2081 31459328-(2080*240*63)=9728 9728/63=154 9728-154*63=26 Geometry with 240 heads 2080/154/26 But: (0*255*63)+(32*63)+33-1= 2048 <-start LBA 2048+31457280= 31459328 <- End LBA 31459328/255/63=1958 31459328-(1958*255*63)=4058 4058/63=64 4058-64*63=26 Geometry with 255heads 1958/64/26 jaclaz

You have an "added difficulty". One of the two systems has a "non-standard" 240 heads geometry. As an example this is typical of a Lenovo laptop (but many "HP" desktops do the same). If you connect a hard disk internally, the BIOS "assumes" or "forces" the CHS geometry to be n/240/63 (whilst the "real" hd geometry is m/255/63. Since the disk is originally partitioned/formatted on that motherboard, the partition table and PBR/bootsector BPB will reflect the 240 heads geometry. When you connect the same hard disk to another "normal" motherboard (or even to the same motherboard but through an USB enclosure/interface) the geometry returns to a 255 heads one and you have problems in: booting running disk utilities (or viceversa, original was 255 and "current BIOS" reads 240, this seems like the case at hand). TESTDISK has an option to adjust geometry (perceived by TESTDISK). Right now it seems like: the current NTFS BPB has a 255 one the partition table (or the hard disk BIOS data) are 240. Try checking the geometry TESTDISK currently uses and change it to the "other" one, in your case the disk is seen by testdisk as 240 whilst it should be set to 255.. Then re-run the scan for partitions. For the record: The partition labeled "PQSERVICE" is usually a recovery partition, typical of most "BIG" OEM's like HP; COmpaq, Dell, etc., if I recall correctly "PQSERVICE" is used by Acer, Gateway and Packard Bell, but many other makers may use it) The "E" is the Extended partition. The "L" is the Logical volume inside it (with label "[DATA]). Since you don't need them, if you don't know what an Extended partiton is, it's allright. jaclaz

NOone is telling you to load the .iso in memory, we are all suggesting you to map the .iso directly, i.e. WITHOUT using --mem mapping. jaclaz

Do you think we are magicians or what? You cannot load an image in memory if you haven't enough of it. The scope of this app is to Install Windows from USB (and NOT to add to it every possible kind of OS, and NOT to twealk the latter to work -if possible - on underpowered machines) You may have better luck with Sardu or Xboot (or any of the other good LInux utilities): http://reboot.pro/forum/100/ http://reboot.pro/forum/126/ or do your own experiments, since this app already installs grub4dos, add the whatever you want manually, after WinsetupfromUSB has run: http://reboot.pro/forum/66/ Some examples of known working OS's: http://reboot.pro/5041/ Remember that if you map a .iso directly the .iso needs to be contiguous. jaclaz

That is NOT a problem, add: Still waiting for the "fake" problematic .inf .... jaclaz

Yes, the ISO standard allows for "links". Basically a "same file" (same name and same size, and probably also same CRC32) if found in the source tree more than once is written to the .iso as follows: first occurrence is the actual file every other occurrence is a link to the first file Since a link only occupies a few bytes, this allows for having all duplicate files fit in roughly half the size (or 1/3 if triplicated, etc.). I am not familiar with Easyboot, you may want to review it's help /docs or ask on it's forum: http://www.ezbsystems.com/cgi-bin/ikonboard.cgi jaclaz

No, meaning that it is NOT confirmed (nor will be). For *any* reason the partition table (or something else) has gone or is invalid, you need to run recovery oriented software to first thing try rebuilding the parition table and filesystem and, should this not be possible, use file-based recovery. A similar/related thread: If you need help for the recovery start a NEW thread. jaclaz

You can use ISO links (or graft-points) or "duplicates once". It depends on which software you use to build the .iso. Some reference: http://reboot.pro/9446/ jaclaz

Yes/no. As I see it it is out of "current scope". All that seems to me like something that may be "post-processed". (and consequently to be tackled later, ONCE we have "basic" functionalities working). Current scope (as I see it) is to "consolidate" and "merge" whatever is in the "source" files into a "monolithic" file with NO CHANGES to the contents WHATSOEVER (apart removing crazy blank spaces and [TAB]'s). jaclaz

Experiment #1 for today : get mkdosfs: http://www1.mager.org/mkdosfs/ http://www1.mager.org/mkdosfs/mkdosfs.zip get dsfok: http://members.ozemail.com.au/~nulifetv/freezip/freeware/ http://members.ozemail.com.au/~nulifetv/freezip/freeware/dsfok.zip get the attached file (usual half-@§§ed batch ) expand everything in a directory run the batch (FAT12build.cmd) Experiment #2 : Try using this app instead of Free Iso Creator: http://www.nbxsoft.com/ Free Create-Burn ISO http://www.nbxsoft.com/files/createburniso.exe It should avaoid patching the 0x02 to 0x03 jaclaz FAT12build.zip

No, the idea behind the two little apps is: no command line parameteres, simply do what is supposed to do. jaclaz

As I see it, right now the issue is to understand which bootsector CODE works with the sfloppy approach and which don't (and if they can tweaked to work). As briefly mentioned in the .xls, the MS-DOS 7.x boot code works, the FreeDOS and the grub4dos one do not (but I did just a quick test, that ened to be repeated/reproduced ). @dencorso generally speaking editing .iso's is not the best idea in the world, I have seen more trouble coming form doing that than almost anything else, the "right way" is to rebuild the .iso, IMHO. jaclaz

Well, no. That's called "I have something that is not currently available, for free or for $. Additionally, WHEN and IF it will be available, it will be "Closed-Source"." Corollary: "I won't also not reveal that I have this thing ready in the closet, unless someone will torture me" Sure, it is something that is done once one has decided to market something, in order to forecast revenues, if they are calculated to be too low, the item is not marketed after all. Of course you are perfectly free to release (or NOT release) anything as well as market (or avoid marketing) any of your tools, that's the very good thing about freedom . But you have to see it from a purely pragmatical viewpoint. Real, physical things produced have usually 5 main source for costs: development costs marketing/advertising costs production costs distribution costs product support costs When you talk about "immaterial" things like software item #3 is 0, item #4 is near to 0 (electronic dowmload) and #5 should be near to 0 as well (if the product is good ). So you have only two items: development costs marketing/advertising costs of which you ALREADY sustained the biggest item #1, so every cent you can get from the product is better than nothing. Pre-marketing surveys may be useful to target the retail price and to validate the usefulness of an advertising campaign, for which till now you had no costs. @dencorso You mean you doubted about it? You deserve to go behind the green glass door as "leery, but heedful peep "! jaclaz

I have used in the past one of the "bidirectional-type", which have a different "shape", like these: http://www.geeks.com/details.asp?invtid=RXD639-IDE-SATA-ADP&cpc=SCH it was just for "emergency", but it did work with 80-pin cables and "acceptable" speed (didn't measure it, at the time, but I would have noticed an ATA-33 speed). Cannot really say which chip it was based. BTW, and OT, IMHO a little lead in an adapter: is UNlikely to provoke *anything* , after all the human race did not became extinct notwithstanding saturnism (which needs quite a bigger amount of lead to be absorbed): http://medical-dictionary.thefreedictionary.com/saturnism @dencorso the spif223a is rated as being capable of haveing full ATA speed: http://www.sybausa.com/productInfo.php?iid=372 http://www.sunplusit.com/english/products/storage/SPIF223A.aspx uploadpdf.ic37.com/2009-2-25/SPIF223A_www.ic37.com.pdf jaclaz

Yes/no. What you are saying is mostly about re-formatting a 720 kb floppy. And yes, XP does attempt reading the existing info on the floppy. JFYI : http://www.winimage.info/forum/viewtopic.php?f=3&t=3117 and yes, NOT "news" : It is perfectly possible to format also a "bulk" floppy as 720 K: OT, but still FYI, besides FORMAT144 we also have FORMAT720: http://www.denispetrov.com/?page_id=3 jaclaz

@rloew Yes, it is very possible, I am usually very accurate and punctilious . Sure they are the headers from the Menuet Os guys, in their own words: http://www.menuetos.net/cdboot.htm more generally I defined myself the approach as half-@§§ed (actually 2/4-@§§ed) exactly to save other people the need of coming out saying: You share some common trait with one of the greatest geniuses in the whole history of mathematics, Pierre de Fermat : @dencorso Experiment for the day : populate your image with MSDOS 7.1 files+whatever else you might want to test use this app: http://www.minidvdsoft.com/isocreator/ to create a bootable .iso with it (+ any files you want in the "accessible CD" part) hexedit in the .iso AA5555AA8802 to AA5555AA8803 try booting the .iso in Qemu or whatever Vm you use jaclaz

Nice find! Possibly an evolution of these thingies here: http://reboot.pro/3537/ http://reboot.pro/3791/ Here lies the problem, there is no "we" , personally I ALREADY won, since more than ten years, as well as - I am told - several satisfied Win2K users, you seem like the only one having these problems (or considering the whatever you have as problems or refusing to follow simple troubleshooting steps or insisting in attributing these problems, whatever they are, to strange conspiracies). You are probably trying to do all together something that you never did, with a non-standard setup, adding to it each and every possible kind of things that may cause the problems. The fact that you come not (like most satisfied Win2K users) from a NT 4.0 background may play a part in some of the issues you describe, though, expecially those related to Admin or "plain user" authorizations and privileges. I was trying, in an as nice as possible way , to send the message that there is a non-trivial amount of probabilities that you managed to create a somehow "botched" install, by either implementing "queer" settings or installing "bad" apps. Keep also in mind that there is very little need, with a decently sized system partition and for a single user, to have the system partition NTFS formatted, Once again, my personal advice: describe your hard disk partitioning/settings AND the actual hardware you are using FORGET about ANY app just make a new, plain, simple, "naked", install of 2K, DO NOT install ANY other software try - next time - to just answer the questions other members may ask (they are intended to better understand the situation/problems and to try and help you) try - still next time - to just follow the advice other (more experienced with 2K) members may give you, once they will understand thanks to the above, WHAT the problem may be report what happens after following the above advice (and doing nothing else) jaclaz P.S.: On the new, "naked" install try following this: http://www.petri.co.il/pagefile_optimization.htm the 1.5x value may be nonsense (depending on the amount of RAM you have) Just in case: http://support.microsoft.com/kb/255205/en-us

Corollary: Anything that works is better than anything that doesn't. Truism or tautology? http://en.wikipedia.org/wiki/Truism http://en.wikipedia.org/wiki/Tautology_(rhetoric) jaclaz

And how did you create the .iso (or CD)? @rloew You need a sturdier desk! (or a self-standing "for sale" sign ) @all Find attached an early, as usual half- @§§ed (in this particular case 3/4 ) approach . Credits go to rloew and to the good guys at MenuetOS : http://www.menuetos.net/cdboot.htm jaclaz sfloppyCD_01.zip

I give up, I can see when it is of no use. From the amount of troubles you describe I must have been lucky (and have been so for the last 11 years) since I never experienced any of the problems you seem like having. BTW on most machines I ever ran and on actual machines that I do still run 2K on, I use OpenOffice also. jaclaz

Strangely enough, the idea is not really-really new http://bootcd.narod.ru/images_e.htm (though in those cases the .rar was inside the .zip, and you made it the other way round ) I won't even try , at least for the moment , I mean, let's go in steps, I would be happy enough to have it work with a 3840 Kb or 5760 Kb one jaclaz

Or a "real-real" one, like Ultradefrag.... http://ultradefrag.sourceforge.net/ Try checking the "status" column in it's report. jaclaz

Who knows? My remark was more a "generic" one, you have something in your closet, you may be willing to sell it, but until you don't take it out of the closet and put it on display on your desk, under a big "for sale" sign you have 100% possibilities (read as "certainty") that noone will ever buy it, or the other way round 0% probabilities of ever selling it. Once you have it in plain view on the desk it is possible that someone is interested to it, you will have n% probabilities that someone will buy it, and no matter how little n will be it will always verify the n>=0 condition, with a chance of also verifying the n>0 one. jaclaz

Not here , using on the above file splitinf.cmd dated 16/07/2011, I can find no "ECHO" in any of the files in the "SPLIT_" directory. Nor, after having used join_dedupe_inf.cmd dated 15/07/2011 in the "JOINED_" .inf. BUT there was a problem (NOT related to the ECHO) in beautify.cmd (forgot a few x's, my bad ) Find atttached the SAME splitinf.cmd and join_dedupe_inf.cmd with a corrected beautify.cmd Try again. jaclaz split_inf_4.zip