Wednesday, November 27, 2002

Taming the Infernal Machine #1

MS Office & the RVP mystery

The purpose of this and the following articles is to provide a few (hopefully helpful) answers to a number of questions troubling the users of personal computers. Let it be made clear from the very beginning that I claim to be no qualified technician, nor comfortable enough with the more complex workings of the Infernal Machine so as to assume responsibility for any misunderstanding or erroneous use of the information given below.

IMPORTANT NOTE: The following article was written more than 10 years ago and concerns now obsolete versions of Microsoft Windows, Internet Explorer, Office and Outlook Express, so most of the advice provided should be considered as inapplicable to modern systems and had better not be followed. It only remains published here for archival and/or "historical" purposes.

My aim in this particular article is to address a number of issues, with primary focus on the specific relation between a certain feature of MS Office 2000 and the functional peculiarities of MS Internet Explorer's caching attributes.

Those of us using MS Office 2000 have most probably noticed that this program is implicitly linked to our browser's cache, enabling us, through a specific command from inside any Word document, to use any of the entries therein in order to place hyperlinks in our documents (of which Office itself is inexplicably proud, on the grounds that it saves us the trouble of manually typing the hyperlinks). On the other hand, most of us are probably also aware that we hardly ever need to insert any of those entries in our documents. Most of them, anyway, usually disappear after cleaning the browser's cache, so there is a safe bet this particular feature is practically of no use.

So whose brilliant idea was it to put this particular feature there, and why?

Well. The key to the answer is contained in the very question. As I mentioned above, most of these entries will go away by clearing the cache. Which means there are some, if only a few, that will stay there no matter how thoroughly you erase the traces of your meanderings through cyberspace, using the legitimate tools provided by the operating system itself, as those usually happen to be the only ones at hand and the safest to use.

The naked, and almost always bitter, truth is that most people do not actually worry about this kind of paradoxes and just let things be, enjoying their computer's wondrous ability to remember passwords, complete half-typed URLs and keep a zealous track record of whatever they do. They even go as far as waxing poetic about the trouble and time it spares them, blissfully oblivious to the fact that there is a reason behind everything in this universe.

Well, maybe this is what makes the difference between a common mortal and a scientifically inclined mind, a normal person and a control freak - or, simply, someone who is not curious about how things work and someone who is.

As a person belonging in this last category, and with a pointed inclination towards analytical thinking, I usually try to get to the bottom of things and work out solutions to apparently unanswerable problems. Perfectionism, hypochondria, or even vanity - whatever you prefer to call it, the point is that in this particular instance, I was puzzled to discover there was no way of deleting the Recently Viewed Pages (RVP) list from the Hyperlink box in Word; only to realize, shortly after, that deletion was impossible, simply because they were located in a file that did not exist! I tried every possible search method, showing all files and looking for text in all file types (which took about an afternoon to complete), scouring the registry and even dismantling a few DLLs. Nothing! I went online to all major search engines and typed and retyped my request, phrasing it in every possible way, and then patiently checked all 5,000 and more search results. Still nothing. Everyone was happily explaining in exhaustive detail how this amazing new feature of Word 2000 would facilitate the insertion of URLs in a document, but no mention of how to get rid of them - under the assumption, I believe, that no one would ever have reasons to consider getting rid of them.

As for my humble individual, the truth is I did not have any viable reasons. The only entries in the infamous box were the error page from a certain server, a search directory pointing to various image and graphics archives and a movie review web site. Nothing to hide, nothing to be ashamed of or even just worry about, in case someone else took a peek. Still, it bugged me to no end that I was unable to delete them. I knew, just knew there was a way, and was determined to find out, even if it meant taking the computer apart.

Another exhaustive search on the Internet brought to light ONE single request by someone who had more or less the same problem. He was looking for answers on a PC troubleshooting forum, only to receive a few vague directions from the assigned tech about reducing the number of days the visited pages should be kept in History, which, of course, was no great help - I had long ago reduced the number of days to 1, and the URLs in Word's Hyperlink box were already more than half a year old.

However unhelpful, though, the tech's reply set the gears spinning in my head. Now I was sure beyond any doubt that the whole issue had something to do with the browser's cache and the cookies. I suspected there was some secret storage space for those indelible cookies that kept the entries appearing in the Hyperlink box. If there was a way to make the invisible files appear, the mystery would automatically be solved.

I tried everything I could think of. I even downloaded the 30-day trial version of a much-praised program which is supposed, among other things, to purge your hard drive from any traces of your Internet activity. I ran it a couple of times, and while it did free some much-needed space on my disk and even overwrote all Downloaded Program Files (which finally proved to be not such a great idea - but that is another sad story), when I opened a Word document and, trembling with anxious hope, clicked the Insert Hyperlink button - yep, you guessed right! The Ominous Entries were still there. Undead. Zombified. One of them, actually, was going on leading its vampiric life even after the respective web site itself had vanished from the face of the Net.

Then, by almost eerie coincidence, I stumbled upon this article. And everything clicked into place.

Based on that, and some further reading and online searching I did, I finally put two and two together and reached the (now obvious and inevitable) conclusion that the indestructible entries owe their bothersome existence to the fruit of some evil machination, called "persistent" cookies, that reside in some dark shadowy place on your hard drive, so well-hidden it passes for non-existent and impossible to locate by any conventional means, playing hide-and-seek with your (virtual) peace of mind.

Now the question is, why in the world would an error page, of all web pages, leave a persistent cookie on your system - and it is one I am neither entitled, nor in position to answer. What I will venture to explain, however, is the computer's annoying habit of spying on you and recording each and every URL you have visited. It already bugs me that my Internet activity, however innocent, should be continually monitored by the various, both known and unknown, cyberspace spyware for reasons well beyond my comprehension (and it is not my purpose to get into that; although I am acutely aware and solidly convinced about the vital importance of such an issue, I will not start ranting about universal conspiracies against individual privacy and the like, because this is not my point right now). But what interest would anyone have in the sophisticated records my PC keeps, since I am the only one using (and often abusing) it? I seriously doubt the thousands of background textures, animated graphics, funny sound files and PC maintenance freeware programs I have so far downloaded from the Web would be of any use to whatever official establishment or secret agency, or that anybody would waste their precious time over the crossword puzzles and cartoons I design for a children's magazine, my half-finished translations of discouragingly bulky novels, or the random literary endeavors spurred on by my twisted Muse.

The answer to all of the above may sound simplistic, but it is the only satisfactory one I could come up with so far: Windows HAS NOT BEEN DESIGNED FOR PERSONAL USE. So the term "Personal Computer" has more or less ended up being reduced to an oxymoron, despite the ever growing number of machines actually operated by home users (most of whom are not necessarily field-trained, so to speak - or even qualified, for that matter). Companies and organizations relying heavily on the immaculate function of their computer networks are, to a certain extent, entitled to monitor the surfing activities of their personnel so as to keep their interior mechanism in good working order and their outward prestige intact. I will not get into any details here - I am sure you get both the picture and my point. Which inevitably leads back to the initial question: is there a way for us independent PC users, who are not part of any network nor have anything to do with whatever third party, to safely clear the machine's secret records - if for no other reason than maintaining a reasonable amount of unoccupied disk space?

Well, yes. There is a way, but it can only be put to effective use by deleting the "persistent" cookies dropped by various sites and for many different reasons, and permanently residing in the infamous INDEX.DAT files which happen to be hiding in the depths of our PC with almost labyrinthine complexity - well, almost, because nothing remains secret under the sun.

The bad news is you cannot delete any INDEX.DAT file from within Windows, because they are supposed to be in use while Windows is running. Fortunately, there is a workaround, if somewhat complicated - but the result is well worth the effort. Which now takes us to the good news: You can truly, permanently and quite reliably purge your browser's cache through DOS. However, I consider it my duty to remind that I hold no responsibility over the results of this process, nor can I guarantee I will be in position to help reverse any potential damage, so if you are not really sure about what you are getting into, simply refrain and let things be.

Note that the following directions are exclusively applicable to MS Windows 9x including 98SE (I tend to separate Windows 98SE from the rest of Windows 9x, because, in my experience, 98SE seems to be a rather "idiosyncratic" platform, much more related to the newest versions of Windows than to 9x), 2000, ME & XP, using MS Internet Explorer 4.0 and above.

To enter Real DOS mode:

Shut down the PC, wait for at least 30 seconds (to let the memory clear) and power on again, holding down the Ctrl or F8 key (if you are using Windows 2000, ME or XP you will need to boot using either a boot floppy you have already created or borrowed from a friend, or the Windows CD).

A numbered menu will appear on the black screen. Using the Down arrow, select Command Prompt Only (selection #5 in most computers), press Enter and then let the process run on its own until you get the final C:\> prompt, with the blinking underscore next to it.

Now type the following commands that appear in bold characters, pressing the ENTER or Return key after each one (every new numbered line corresponds to a new prompt), and of course, do NOT include the list numbers or the explanations given between brackets:

1. cd windows (this command takes you to the Windows directory, which means you get a C:\WINDOWS> prompt with a blinking underscore next to it)
2. smartdrv (where Smartdrv is a program that will help speed up the process. Wait for it to fully load until you get another C:\WINDOWS> prompt with a blinking underscore, before typing the next command)
3. cd\ (this command takes you to the contents of your C drive, which means you get a C:\> prompt with a blinking underscore next to it)
4. deltree/y temp (the y "switch" saves you from manually confirming the deletion of every file in the directory)
5. cd windows (takes you back to the Windows directory, giving you once again the C:\WINDOWS> prompt with the blinking underscore)
6. deltree/y cookies (deletes the Cookies folder)
7. deltree/y temp (deletes the Temp folder)
8. deltree/y history (deletes the History folder)
9. deltree/y tempor~1 (deletes the Temporary Internet Files folder)

To return to Windows, type win or exit then press ENTER or Return, and you're done.

There is absolutely no need to worry about deleting the above folders. Like the Phoenix, they have the magical attribute of being reborn from their ashes every time you reboot your PC, so your system should suffer no loss. The only difference is that they will now be squeaky clean, having been regenerated from scratch, and you will most probably notice a considerable alleviation in terms of hard disk space.

Once your PC has re-entered Windows mode, open any Word document, click Insert Hyperlink and then Recently Viewed Pages. See? All clear. Don't tell me it wasn't worth the while!

ATTENTION: In some systems, especially for the latest Windows versions, Temporary Internet Files may be hiding in some place other than Windows, or in a folder within the Windows folder, e.g. Application Data (and perhaps even a hidden one, for that matter). From Explorer's Options (or View, or Tools, depending on the platform) menu, select Show All Files and then run Find File to locate INDEX.DAT. There should be two of them, one in the Temporary Internet Files folder and the other in History folder. Write down the location of these two folders, as well as their MS-DOS names (you should see those if you highlight each folder, right-click and choose Properties). In Real DOS mode, type the complete address for each at the prompt, beginning with the directory (e.g. applic~1) at the C:\WINDOWS> prompt, and ending with the folder containing the one you want to delete. Remember to separate them with an inverted slash (\) and NOT USE SPACES except where indicated, or you'll get a Too Many Parameters error message. Then type deltree/y followed by a space and the MS-DOS name of the folder you want to delete, e.g. tempor~1 and press ENTER or Return at the end of the command. Repeat the exact process with the Temp, History and Cookies directories.

No comments: