Wild -- I think the previous poster must have hit it on the head with his observation about phonetic vs. shape-recognition reading style, and I must do the latter. It took me three or four times as long to read the mangled original as to read the correctly spelled version. I got it, but I had to stop and sound it out in my head, which I more or less never do when I'm reading (well-edited) text.
Spelling errors usually grate on me, though if they're occasional and minor I don't bother complaining. We all make mistakes and our fingers are all just slightly too thick for the keys sometimes. I've forced myself to add common misspellings to my visual dictionary, like "theif" vs. "thief."
But for some reason misspelled words stand out from the page as if they're bold, italic, and <blink>ing, and when I'm reading or skimming I almost always come to a dead stop when I hit one. It's kind of a sudden "something here isn't right" feeling. Severely mangled grammar gets me too, though less reliably.
And yes, sometimes when I'm reading newsgroups or message boards it's a brain feature I really wish I could switch off!
It's not all bad, though: when I was in school my friends used me as a high-speed term-paper proofreading machine. Five seconds of scanning a typical typewritten page and I'd've spotted most of the spelling errors and a lot of the grammar errors. And I don't think I ever once had to have one of my papers checked for spelling and grammar, a big time-saver in the days before decent automated spell checking.
Slashdot could use a few editors with this skill/curse, I think.
To bring this back on-topic... I wonder how strict the gloves in question will be about spelling (of words built out of signs for letters) or exact finger and hand positions. Seems like you'd want them to be as generous in that regard as possible, which probably makes the problem that much harder to solve.
Yes! I was interviewing candidates for a senior database developer position a few companies back, and it was amazing to me how many people with "10 years SQL experience!" on their resumes couldn't answer simple questions like that.
One of my favorites, which was a great weeder-out, was, "What's the difference between where and having?" Not something I'd expect someone to know if they'd just done a few small database apps on the side, but it shouldn't be obscure for a hard-core database programmer with experience on large projects. Like the parent poster, I found maybe 1 out of 10 candidates could answer it correctly.
In C terms, this would be like asking, "What's the difference between void *foo and void (*foo)()?" Not something a junior developer might need to know, but you just have to laugh at someone who bills themselves as an ace C hacker and can't give you at least a ballpark answer.
Along similar lines, for senior Java programmers, I like "What's the difference between transient and volatile?" Though that one is more of a "bonus points for getting it right" than an "interview's over if you can't answer" for most positions since you can write a huge, well-architected Java app without ever needing either of those keywords. (But if they can answer that and tell me what the strictfp keyword means, I can be pretty sure they know their way around a Java compiler.)
Of course, as others have said, little quiz questions like those are almost always less important than finding out about someone's work habits, personality, coding style, and so forth. Most often I spend at least 75% of an interview on those softer topics rather than on technical stuff.
No, it's not a democracy, and that's fine by me. The alternative is mob rule -- "whatever the majority thinks, goes" means any minority that the majority doesn't like is pretty much screwed. Regardless of the merits of this particular case, I think it's desirable for a government to protect wronged parties from the whims of the majority.
And you're not the only one writing to your representatives about this, though I doubt many others are doing it daily.
If there was no advertising, there would be no commercial need for TV in the US, not even PBS.
What about HBO, Showtime, and the like? A pretty significant percentage of what I watch is on those two networks, and I can guarantee you their CEOs don't care what I do with my fast forward button as long as I remain a paying subscriber.
No question that television advertising is an important influence on culture, but there's a proven business model for TV without ads. (Not only that, the shows on pay channels seem to be better on average, though of course that's a matter of personal taste.)
I do almost the same thing, with one tweak: the procmail script that decides what to do with the mail does some keyword/pattern scanning as well as running Vipul's Razor, and it will only push the mail over to my "require a reply" script if it looks like it might be spam.
The advantage is that I don't have to worry nearly as much about false positives on my spam filters; this system makes them much less expensive than they'd be if I simply tossed all matching mail to/dev/null. It successfully filters out nearly 100% of my spam (75-100 spams a day, of which one or two a week get through.)
I've had this system in place for a year or so, and in that time, maybe 1 out of 25 legitimate personal messages from unknown senders has required a validation E-mail, so it's not a major inconvenience for a huge number of people.
Not that I think it's much of an inconvenience anyway, and I have yet to get one complaint. In fact, Slashdot is the only place I've heard anyone complain about it. The comments I've gotten from actual correspondents have been more along the lines of, "I get too much spam too! How do I set up the same thing on my mailbox?"
I suppose this is going a bit off topic, but how do you get TMPGEnc to do lossless editing? It's easy enough to get it to re-encode the whole video, but I haven't seen any options to let you pass the original compressed data through unmodified.
Didn't see this addressed in the documentation, so maybe someone here knows: Will Cinelerra edit MPEG-2 program streams without reencoding the audio and video? It'd be swell to be able to take the MPEG-2 encoded video from a ReplayTV or TiVo, clip out commercials, and burn to a DVD, but the trick is to do it without reencoding (which would cause quality loss.) Obviously the software would have to generate new keyframes in a few places depending on where the edit points were, but it ought to be able to copy most of the stream without modification.
The only software I've found that does this is M2-Edit
by MediaWare Solutions, but its UI is awful and it's Windows-only.
Like a lot of other people, I use temporary E-mail addresses when I post to Usenet and I give out unique addresses to all the Web sites I register with, so I can easily shut off a particular address if it gets sold to spammers. But the mail still gets delivered to my server, eating bandwidth and CPU time.
It occurred to me that it ought to be possible to use DNS to stop blocked mail from even making it to my server. If, instead of myname-uniquetag@foo.com (I'm using qmail, so the "-uniquetag" is a personal alias I can control without becoming root), I could supply an address like myname@uniquetag.nospam.foo.com, it'd allow me to get rid of the MX record for uniquetag.nospam.foo.com if that address started getting spammed. Presto, spam doesn't know where to go and doesn't fly across the net eating resources. Granted, I'd get the DNS requests, but that's true today anyway.
Seems like all this would require would be a few simple tools to let users add and remove subdomains on the fly.
Not a replacement for filtering by any means (you still have to deal with the more common case where an address gets a mix of legit and spam messages, and it's hard to imagine a non-techie user ever using this) but it's another possible weapon in the anti-spam arsenal, one that attacks the waste-of-bandwidth problem.
It's definitely a concern for some people! At a previous job I wrote wrappers for all the time-related C library calls our code used, and made the application code use 64-bit time everywhere. Nowadays I mostly code in Java, which uses 64-bit time from the get-go (Java time is in milliseconds instead of seconds, but that's still a lot of headroom.)
That said, I agree with the parent that there seems to be much less concern about the problem than there ought to be. The crazy thing is how long it's taking OS vendors to supply low-level 64-bit time system calls. If I could have used 64-bit time in my stat() calls and so forth, I would have started doing it years ago. But short of not looking at the clock or writing wrappers like I did, it's impossible to code a Y2038-proof application under some OSes even today, and on the OSes where it is possible, it usually takes some hunting to figure out how. Most vendors have tweaked their system calls to allow 64-bit file sizes, but for some incomprehensible reason they didn't move to larger time values while they were breaking the APIs anyway.
Time representation is one place I think Microsoft got it right, actually. One of the Windows time formats is a floating-point value, the number of days since Jan. 1, 1900 if I recall correctly. This is great since it gives you sub-microsecond precision for the immediate future while allowing dates way off in the past or future.
Three cheers to Shuttle for taking noise into account in their design, but I hardly think it's reasonable to describe 51dB as "near-silent" like the review does. With careful selection of components and some custom tweaks (replacing power supply fans, etc.) it's not too outrageously difficult to get a reasonably powered PC in the mid-30dB range. My desktop PC is about 35dB and it's still louder than I'd like.
If you're willing to go with water cooling or something like the CALM System's evaporative cooling, you can get even quieter than that. Or you can do what I did and run cables under the floor so you can put the thing in a closet and shut the door.
Given the link between sleep and memory (short-term memory is thought by many researchers to transfer to long-term during sleep, for example) I wonder if there'll be a similar effect among users of modafinil
and other sleep-suppression drugs. Or is it something unique to caffeine that has nothing to do with its effect on sleep cycles?
Now, tell me why you're worth that hassle to me.. Why I shouldn't treat your words with the same respect you treat mine, and bitbucket them.
Not that I suspect there's any hope of either of us convincing the other here, but... I don't bitbucket your message. Your original mail gets saved and delivered when you reply to the autoresponse.
But that aside, I actually agree with you! Maybe I'm not worth the time it takes you to deal with my autoresponder. It all depends on what you were sending me and why. If someone who's not already on my whitelist sends me a message that's worth that little to them, chances are it won't be worth too much to me either.
I should note that they also won't get an autoresponse if their message scores low enough on my looks-like-spam tests, which in practice lets the majority of mail from unknown people through immediately. That was why I wrote my software, as an answer to the fact that it's impossible to aggressively detect spam without getting false positives. With the autoresponder, the cost of false positives is very low and I can afford to make my detection broad enough to catch just about all the spam I get. So if your initial message looks nothing like spam to my filters, you'll never know I have an autoresponder.
For what it's worth, I'd consider it a worthwhile tradeoff if everyone I communicated with were running this kind of software even without the passthrough for legit-looking messages. Certainly much less inconvenient and time-consuming than the spam it'd prevent -- more like everyone running yellow lights than red ones.
Also, how would you like to, every other time you sent an email, have to handle a braindead acknowledgement.
Only if you never send mail to the same person twice. Once you're on someone's whitelist you don't see another acknowledgement request.
I use a similar setup (custom-written) and it does a wonderful job of cutting down on spam. I have yet to get a reply from someone who found it a significant inconvenience -- on the contrary, the most frequent comment I get about it is, "Wow, how can I do the same thing with *my* mail? I get way too much spam!"
To answer your question, mailing lists that I know I'm on get to bypass the acknowledgement filter, but their mail still gets run through other filters (Vipul's Razor, etc.) which catch most of the spam people send to them. Using qmail, I can also give people a unique private address that bypasses some or all of my filters but that I can shut off completely without affecting anyone else.
So in practice, just one more tool in the toolkit, but it catches a good 74 out of the ~75 spam messages I get each day, and as far as I can tell has yet to cost me a single legit message.
I discovered after downloading and building 3.4p1 on my Solaris 7 box that Solaris doesn't support a shared memory feature 3.4's privilege separation uses. As a result, you can enable privilege separation or compression, but not both at the same time. Just something to be aware of if you're considering an upgrade. (It's possible more recent versions of Solaris don't have this problem.)
What does this have to do with Microsoft? Reducing support costs is an industrywide -- no, make that societywide -- trend. Microsoft may be doing the same thing, but to my mind it's ludicrous to blame them for other companies' perception of tech support as an expense they need to reduce to be (more) profitable. In what sense do you think Microsoft is behind it?
That might hold water if the typical car thief stole just one car in his lifetime. I really doubt that's the case.
Me, I'd consider it harmful to have my car stolen, and if the police can reduce the likelihood of such without violating anyone's civil rights (and I don't believe there's a right to steal cars, whether or not the police own them) I'm all for it.
The presence of bait cars may wind up being a strong deterrent to car thieves, since they'll be unable to tell whether a given car will get them caught or not. Even if you say you only want police to act as deterrents, you have to agree that no deterrent is effective unless it bites once in a while.
And exactly how much money do you think they're making from the guy in the article?
I've done something similar (but simpler) on my mail server; I modified qmail to add a 90-second delay before it spits out its "can't send to that address" error. Doesn't stop spammers from scanning me, but hopefully it makes their job more expensive. The delay is hopefully short enough that legitimate clients (if there are any legitimate clients that try to relay when they shouldn't) won't completely fail.
The tweak is a really easy one-liner, BTW; just add a sleep call to err_nogateway() near the top of qmail_smtpd.c.
"Very few" uses, maybe, but there's at least one legitimate use: it allows you to turn off the encryption 2.5 normally applies to video files, a necessary step if you want to archive recordings on a digital medium (DVD-R, for example) without going through a lossy D-A-D conversion.
That's all I plan to use it for. I've already paid the lifetime subscription fee for my DTivo (so I couldn't use it to steal TiVo service even if I wanted to) and I'm happy to pay DirecTV every month for the channels I receive. I have no interest whatsoever in extracting shows to give or sell to other people. I just want to stop accumulating videotapes and start accumulating DVD-Rs instead.
While I think it's great to see networking treated as a public utility, that seems like an awfully long time for a single deployment project. Unless they're planning to update their technologies over time, I don't imagine the last few people getting hooked up in 2012 will find today's broadband networking all that interesting. Would you really want your city to be rolling out 1992-era network technology this year?
Of course, that's less true for really low-income families; if it's a choice between today's speeds and no connectivity at all, it's a no-brainer.
And it's also perhaps the case that 10 years is the maximum time limit in the RFP; maybe bidders will propose to get the job done a lot faster.
Should be an interesting one to watch, in any case, and I expect other cities will follow suit as more and more government functions migrate to the Web. In fact, it won't surprise me to see attempts to force Web-friendly cities into providing public net infrastructure; if the only way you can access a particular city service is via the city Web site, then one could make the case that the city needs to make sure all its citizens can get at the city Web site. I bet someone will file a suit to that effect at some point.
This reminds me of an article I read in Analog magazine years ago about building massive convection towers along the coast of the US to deflect hurricanes.
According to the article, by messing with the relative temperature and humidity at ground level and at high altitude, you can create a pressure system that isn't conducive to hurricanes, which would tend to essentially push an incoming hurricane off to the side. Build a line of these along a hurricane-prone coast, the author said, and they'd pay for themselves within a decade just from the cost savings of not having to rebuild after a couple major storms.
I was pretty skeptical, but what little I remember about the principles behind the idea sounds almost identical to what today's article describes, so maybe it wasn't as kooky as it sounded.
Did anyone else read the same Analog article? Please post if you know which issue it was in -- I'd love to go reread it, since I've forgotten most of the technical details at this point.
Slashdot Mirror
Spelling errors usually grate on me, though if they're occasional and minor I don't bother complaining. We all make mistakes and our fingers are all just slightly too thick for the keys sometimes. I've forced myself to add common misspellings to my visual dictionary, like "theif" vs. "thief."
But for some reason misspelled words stand out from the page as if they're bold, italic, and <blink>ing, and when I'm reading or skimming I almost always come to a dead stop when I hit one. It's kind of a sudden "something here isn't right" feeling. Severely mangled grammar gets me too, though less reliably.
And yes, sometimes when I'm reading newsgroups or message boards it's a brain feature I really wish I could switch off!
It's not all bad, though: when I was in school my friends used me as a high-speed term-paper proofreading machine. Five seconds of scanning a typical typewritten page and I'd've spotted most of the spelling errors and a lot of the grammar errors. And I don't think I ever once had to have one of my papers checked for spelling and grammar, a big time-saver in the days before decent automated spell checking.
Slashdot could use a few editors with this skill/curse, I think.
To bring this back on-topic... I wonder how strict the gloves in question will be about spelling (of words built out of signs for letters) or exact finger and hand positions. Seems like you'd want them to be as generous in that regard as possible, which probably makes the problem that much harder to solve.
One of my favorites, which was a great weeder-out, was, "What's the difference between where and having?" Not something I'd expect someone to know if they'd just done a few small database apps on the side, but it shouldn't be obscure for a hard-core database programmer with experience on large projects. Like the parent poster, I found maybe 1 out of 10 candidates could answer it correctly.
In C terms, this would be like asking, "What's the difference between void *foo and void (*foo)()?" Not something a junior developer might need to know, but you just have to laugh at someone who bills themselves as an ace C hacker and can't give you at least a ballpark answer.
Along similar lines, for senior Java programmers, I like "What's the difference between transient and volatile?" Though that one is more of a "bonus points for getting it right" than an "interview's over if you can't answer" for most positions since you can write a huge, well-architected Java app without ever needing either of those keywords. (But if they can answer that and tell me what the strictfp keyword means, I can be pretty sure they know their way around a Java compiler.)
Of course, as others have said, little quiz questions like those are almost always less important than finding out about someone's work habits, personality, coding style, and so forth. Most often I spend at least 75% of an interview on those softer topics rather than on technical stuff.
And you're not the only one writing to your representatives about this, though I doubt many others are doing it daily.
What about HBO, Showtime, and the like? A pretty significant percentage of what I watch is on those two networks, and I can guarantee you their CEOs don't care what I do with my fast forward button as long as I remain a paying subscriber.
No question that television advertising is an important influence on culture, but there's a proven business model for TV without ads. (Not only that, the shows on pay channels seem to be better on average, though of course that's a matter of personal taste.)
The advantage is that I don't have to worry nearly as much about false positives on my spam filters; this system makes them much less expensive than they'd be if I simply tossed all matching mail to /dev/null. It successfully filters out nearly 100% of my spam (75-100 spams a day, of which one or two a week get through.)
I've had this system in place for a year or so, and in that time, maybe 1 out of 25 legitimate personal messages from unknown senders has required a validation E-mail, so it's not a major inconvenience for a huge number of people.
Not that I think it's much of an inconvenience anyway, and I have yet to get one complaint. In fact, Slashdot is the only place I've heard anyone complain about it. The comments I've gotten from actual correspondents have been more along the lines of, "I get too much spam too! How do I set up the same thing on my mailbox?"
I suppose this is going a bit off topic, but how do you get TMPGEnc to do lossless editing? It's easy enough to get it to re-encode the whole video, but I haven't seen any options to let you pass the original compressed data through unmodified.
The only software I've found that does this is M2-Edit by MediaWare Solutions, but its UI is awful and it's Windows-only.
It occurred to me that it ought to be possible to use DNS to stop blocked mail from even making it to my server. If, instead of myname-uniquetag@foo.com (I'm using qmail, so the "-uniquetag" is a personal alias I can control without becoming root), I could supply an address like myname@uniquetag.nospam.foo.com, it'd allow me to get rid of the MX record for uniquetag.nospam.foo.com if that address started getting spammed. Presto, spam doesn't know where to go and doesn't fly across the net eating resources. Granted, I'd get the DNS requests, but that's true today anyway.
Seems like all this would require would be a few simple tools to let users add and remove subdomains on the fly.
Not a replacement for filtering by any means (you still have to deal with the more common case where an address gets a mix of legit and spam messages, and it's hard to imagine a non-techie user ever using this) but it's another possible weapon in the anti-spam arsenal, one that attacks the waste-of-bandwidth problem.
That said, I agree with the parent that there seems to be much less concern about the problem than there ought to be. The crazy thing is how long it's taking OS vendors to supply low-level 64-bit time system calls. If I could have used 64-bit time in my stat() calls and so forth, I would have started doing it years ago. But short of not looking at the clock or writing wrappers like I did, it's impossible to code a Y2038-proof application under some OSes even today, and on the OSes where it is possible, it usually takes some hunting to figure out how. Most vendors have tweaked their system calls to allow 64-bit file sizes, but for some incomprehensible reason they didn't move to larger time values while they were breaking the APIs anyway.
Time representation is one place I think Microsoft got it right, actually. One of the Windows time formats is a floating-point value, the number of days since Jan. 1, 1900 if I recall correctly. This is great since it gives you sub-microsecond precision for the immediate future while allowing dates way off in the past or future.
If you're willing to go with water cooling or something like the CALM System's evaporative cooling, you can get even quieter than that. Or you can do what I did and run cables under the floor so you can put the thing in a closet and shut the door.
Yahoo's Silent-PC list and AVS Forum have lots of good info on building PCs a lot quieter than the new Shuttle.
But not as tiny!
Given the link between sleep and memory (short-term memory is thought by many researchers to transfer to long-term during sleep, for example) I wonder if there'll be a similar effect among users of modafinil and other sleep-suppression drugs. Or is it something unique to caffeine that has nothing to do with its effect on sleep cycles?
Not that I suspect there's any hope of either of us convincing the other here, but... I don't bitbucket your message. Your original mail gets saved and delivered when you reply to the autoresponse.
But that aside, I actually agree with you! Maybe I'm not worth the time it takes you to deal with my autoresponder. It all depends on what you were sending me and why. If someone who's not already on my whitelist sends me a message that's worth that little to them, chances are it won't be worth too much to me either.
I should note that they also won't get an autoresponse if their message scores low enough on my looks-like-spam tests, which in practice lets the majority of mail from unknown people through immediately. That was why I wrote my software, as an answer to the fact that it's impossible to aggressively detect spam without getting false positives. With the autoresponder, the cost of false positives is very low and I can afford to make my detection broad enough to catch just about all the spam I get. So if your initial message looks nothing like spam to my filters, you'll never know I have an autoresponder.
For what it's worth, I'd consider it a worthwhile tradeoff if everyone I communicated with were running this kind of software even without the passthrough for legit-looking messages. Certainly much less inconvenient and time-consuming than the spam it'd prevent -- more like everyone running yellow lights than red ones.
Only if you never send mail to the same person twice. Once you're on someone's whitelist you don't see another acknowledgement request.
I use a similar setup (custom-written) and it does a wonderful job of cutting down on spam. I have yet to get a reply from someone who found it a significant inconvenience -- on the contrary, the most frequent comment I get about it is, "Wow, how can I do the same thing with *my* mail? I get way too much spam!"
To answer your question, mailing lists that I know I'm on get to bypass the acknowledgement filter, but their mail still gets run through other filters (Vipul's Razor, etc.) which catch most of the spam people send to them. Using qmail, I can also give people a unique private address that bypasses some or all of my filters but that I can shut off completely without affecting anyone else.
So in practice, just one more tool in the toolkit, but it catches a good 74 out of the ~75 spam messages I get each day, and as far as I can tell has yet to cost me a single legit message.
I discovered after downloading and building 3.4p1 on my Solaris 7 box that Solaris doesn't support a shared memory feature 3.4's privilege separation uses. As a result, you can enable privilege separation or compression, but not both at the same time. Just something to be aware of if you're considering an upgrade. (It's possible more recent versions of Solaris don't have this problem.)
Reduce your exposure to RF emissions by carrying around a powerful RF transmitter! Sure, that'll do the trick.
What does this have to do with Microsoft? Reducing support costs is an industrywide -- no, make that societywide -- trend. Microsoft may be doing the same thing, but to my mind it's ludicrous to blame them for other companies' perception of tech support as an expense they need to reduce to be (more) profitable. In what sense do you think Microsoft is behind it?
It is quite possible to pull MPEG-2 streams off a TiVo and put them on your PC's hard drive.
Me, I'd consider it harmful to have my car stolen, and if the police can reduce the likelihood of such without violating anyone's civil rights (and I don't believe there's a right to steal cars, whether or not the police own them) I'm all for it.
The presence of bait cars may wind up being a strong deterrent to car thieves, since they'll be unable to tell whether a given car will get them caught or not. Even if you say you only want police to act as deterrents, you have to agree that no deterrent is effective unless it bites once in a while.
And exactly how much money do you think they're making from the guy in the article?
He has a day job. Sun doesn't.
The tweak is a really easy one-liner, BTW; just add a sleep call to err_nogateway() near the top of qmail_smtpd.c.
Reprehensible though they may be, legal bullying tactics work.
That's all I plan to use it for. I've already paid the lifetime subscription fee for my DTivo (so I couldn't use it to steal TiVo service even if I wanted to) and I'm happy to pay DirecTV every month for the channels I receive. I have no interest whatsoever in extracting shows to give or sell to other people. I just want to stop accumulating videotapes and start accumulating DVD-Rs instead.
And I thought lead in the water supply was bad enough. Now I have to have my water tested for excessive robot levels.
While I think it's great to see networking treated as a public utility, that seems like an awfully long time for a single deployment project. Unless they're planning to update their technologies over time, I don't imagine the last few people getting hooked up in 2012 will find today's broadband networking all that interesting. Would you really want your city to be rolling out 1992-era network technology this year? Of course, that's less true for really low-income families; if it's a choice between today's speeds and no connectivity at all, it's a no-brainer. And it's also perhaps the case that 10 years is the maximum time limit in the RFP; maybe bidders will propose to get the job done a lot faster. Should be an interesting one to watch, in any case, and I expect other cities will follow suit as more and more government functions migrate to the Web. In fact, it won't surprise me to see attempts to force Web-friendly cities into providing public net infrastructure; if the only way you can access a particular city service is via the city Web site, then one could make the case that the city needs to make sure all its citizens can get at the city Web site. I bet someone will file a suit to that effect at some point.
According to the article, by messing with the relative temperature and humidity at ground level and at high altitude, you can create a pressure system that isn't conducive to hurricanes, which would tend to essentially push an incoming hurricane off to the side. Build a line of these along a hurricane-prone coast, the author said, and they'd pay for themselves within a decade just from the cost savings of not having to rebuild after a couple major storms.
I was pretty skeptical, but what little I remember about the principles behind the idea sounds almost identical to what today's article describes, so maybe it wasn't as kooky as it sounded.
Did anyone else read the same Analog article? Please post if you know which issue it was in -- I'd love to go reread it, since I've forgotten most of the technical details at this point.