In November of 2000 a guy calling himself John Titor turned up online claiming to be a time-traveller from 2036. Here are a few thoughts on why it's bullshit.
There's a complete archive of his postings and info at johntitor.com
This was a post on a BBS and is therefore a bit less coherent than it would normally be.
My first though was that he is a normal, present-day guy has a stockpile of the IBM 5100s he's looking for and he's going to get rid of them on eBay.
There are several technology-related conflicts in his story:
1) They can build a fucking *time machine* and not replicate either the hardware or functionality of that IBM machine? I think not.
2) Even if that could somehow be the case, he negates it by saying that the "Web" (improperly labeled as people in the future who depend on the Internet for all communications, which is more important as travel is much less feasable would be much more aware that the "Web" is a protocol that runs on a network that has many other uses—it's as stupid as referring to the Internet as "The Email") is much faster and handles all communications. This means that they still have the ability to manufacture and refine processors of all kinds (at least CPUs, DSPs, and comm) and work on advanced software. Given that, they do not need IBM portables from the 1970s.
3) Everyone and their brother knew the basic problem of Y2K. If they need 5100s so much that they're willing to take the risk of and develop the technology to send people back in time to help fix it, 2038 is obviously a huge problem for them and it's only two years away (as he claims tobe from 2036). As a society much more dependant on computers than we are, the chance of someone not understanding the problem (which takes about 20 seconds to grasp), especially someone in the position of being sent away from him family and friends to go to a different fucking time, and at a time when it seems that basic technical literacy would be higher, is nil.
He says: "I'm not exactly sure what the technical issue is but I believe some sort of UNIX system registry stops in 2038." Bull. Shit. ("'unix system registry"??!?)
There's also the basic solution: 64 (or more) bit integers. Just re-define time_t with more bits and recompile. Failing that, it's 2003 and there are 64 bit processors out there and unices using them. As it's considered the next thing and companies are spending billions to develop them, there is good reason to believe that they will be more unbiquitous by 2005 when the war supposedly starts (as unrest, it doesn't become full-blown for several more years). Ergo, no problem.
Even if that's not true, they clearly have processor design and fabrication abilities and working designs would be long-established, so they could fix it themselves. Ergo, no problem.
Or they could have done a software hack (which I've just thought of and there may be problems with, so bear with me), something like modifying the time system calls to use strings behind the scenes, or internally using two time numbers, one for 1970-2038 and one for 2038-whenever the next 32 bit int runs out. Or a binary search and replace that looks through executable code and fixes date types. Something. In any case, going back to an 8 or 16 bit system seems an unlikely solution. Ergo, no problem.
Also, if they knew exactly what would help them about the 5100, they could have replicated it. We're talking about a program that runs on a portable computer from the 1970s, not re-writing a million-plus lines of spaghetti code to replicate XP.
All in all, I think the whole thing is interesting. I think the guy is either a social sciences person just seeing what happens, and generally doing it well, or just some loon.
Heh! time travel;
If it were possible I'd nip back, and ask the Gyn to suppliment my birth mothers Endo system, Problem solved
Posted by Lynda at December 18, 2004 09:20 PMeven 32 bit linux has 64 bit everything, it just does 64 bit things a little slower. I've been slogging through the 64 bit file calls in c, so there is no need for a 64 bit processor before you can implement a 64 bit system call. Heck even the 8080 could do floating point emulation of many bits and it was a lousy 8 bit slowpoke.
Posted by stopmenow at February 3, 2005 01:18 PMI've got a confession to make, dude. I am a time traveller. I came back from the year 2026 a couple of years ago to close some loose ends. You are the one who built the machine. It works. But you're going to spend all your family's fortune on it.
Posted by Snarf at February 6, 2005 04:59 AMAny computer can handle any number of any size, given the considerable memory size.
Just because it's a 32 bit processor has *nothing* to to with the way strings and dates can be stored. as stated before, a 32 bit processor can deal with 64 bit numbers it just takes more operation (it can't to the comparison in one *chunk*).
Any computer can compare the string "aaaaaaaaa" and "aaaaaaab" and compare which one is binary larger. But wait.. 8 bytes is 64 bits! Woah! No Machine could possibly do that!
The problem is, A 16 bit processor can deal with a 2067 bit number fine, it just needs umpteen more cycles.
The reason dates are a problem because the software *stores* it as a 32 bit number, and the unix system calls only really accept 32 bit numbers if compiled on a 32 bit processor. No ones stopping you from creating your own function to deal with an array of 32 bit numbers (there's multiple ways of dealing with HUGENUMBERS on any computer), and storing it as 1208 bits per date or any other arbitrary number.
Now, *that's* the problem.
I firmly believe the entire John Titor thing can be debunked solely from that one date problem.
It cries out wolf that you even need outdated hardware to solve a *software* problem. Software problems are solved by manipulating the code which governs the program, not by needing new PC hardware to figure it out.
I firmly beleieve and stand by the fact that this entire Titor story does not stack up on this evidence alone, and since that UNIX 2038 thing can obviously be proved as a hoax on his part (the 2038 thing WILL happen, don't get me wrong on that, and it will probably cause a little worry here and there, but it will NOT be fixed by some 1970s computer... it will be approached the same way the Y2K thing was.. change as much code as possible in as little time as you had. Again, it's he software approach. You don't need an 70's IBM to fix a software engineering problem.)
Probably there are too many, far to many blogs dedicated to the theoretical physics involved in debunking Titor. That's a shame, as his computer scam doesn't stack up in the first place, without having to bother having to debate the physics.
Posted by Hocks at March 3, 2005 05:10 PMthe IBM 5100 can rewrite the code...
Posted by u at March 12, 2005 12:22 AMIt doesn't strike me as strange that they could create a time machine and not be able to hurdle the 'lesser' technical issues related to the computers. It depends entirely on the knowledge of the survivors and technology, even for highly skilled people, may as well be magic if it is not in their field. The couple of physicists that I know can't even figure out AOL.
Kill off spammers and I'd think the remaining Internet would seem to run faster by they way. Especially if you're on a next generation Internet that's built between then and now.
Which reminds me, if you ask any hump off the street what the difference between the Web and the Internet is, most won't know. Kill off 90% of the people, live in a radioactive environment and lose formal schooling for 20 years, the distinction wouldn't be worth making even if you did. ;)
Posted by Reason at March 26, 2005 08:19 PM