Aside from some dated information and inaccuracies, the problem I've had with the LMD for quite some time is that it's couched as a double negative: "Linux isn't evil foo", where foo is some undesirable characteristic. The entire flavor of the document would change if language were changed from negative to neutral or even positive. Otherwise it has this "are you still beating your wife?" flavor. The LMD would answer that question with a topic "Linux is no longer beating its wife".
Just as an example, the document would have an entirely different flavor if the system headings under "4. Systems Myths" were:
Multitasking
System stability
OS footprint
Ease of networking
System security
Unix "run-alike"
Unix diversity
Linux and innovation
Linux and multiple development tracks
Y2K compliance
Data security
Suggestion is a powerful tool. The document should suggest that Linux is the cat's pyjamas. People will tend to believe.
I do strongly feel that open-plan and private-office workspaces fall at either end of a large continuum, and probably are not ideal for much contemporary technical work. The shared office or high-walled cube approach seems to work better than either open or private layouts.
Of the environments I've worked in, two of the most conducive where a "pod" arrangement of cubes, in which four cubes shared a common corridor access. Scooting back on my chair I could convene with co-workers on similar projects, yet within my cube, I had relative privacy. The other was a "coding pen" with three developers at standard desks pushed up against walls. Small enough group that multi-way communications weren't overly common, enough people that collaboration and quick queries were possible.
One of the worst environments was at Visa International, which has "rationalized" its contractor workspace environment by placing all contractors on a single floor of its Foster City offices, dubbed the UN for its international flavor. My direct reports were six floors above me, and co-worker contractors were a dozen cubes down the aisle and across the floor. Hardly conducive to communications (but I got the skinny on the Y2K project and the marketing design campaign).
Private offices, in a development enviroment, tend to cut too many communications channels. The worst situation is closed doors and no glass -- it's hard to tell what you're walking into and/or whether or not it's appropriate. Even a cube-orientation can be almost as effective. The same office which originally had a "coders pen" moved to a new location in which a high-walled four-workstation unit shaped like a "+" was the initial furnishings. Interactions were virtually stifled: the only ways to communicate were to walk around the entire periphery, or to shout in a voice audible over the entire office. Later, a section of cubes with mixed wall heights was adopted -- communications between common working groups was possible, distractions from other groups were minimized.
With the Internet, the opportunities for distraction have also increased. As you say, downloads, pr0n, computer games, personal business, are all distractions. I don't feel that this is the best mode of operation for most technical work, particularly for people who aren't used to working in such an autonomous environment, unless a well defined (and resonably revisable) set of performance milestones can be established, in which case I really don't care where you work or how you spend your time, so long as you produce.
My current employer has just moved from a large-office, high-walled cube environment to a workspace typified by smaller office spaces (8 to 20 workstations), with low cube walls -- about 5'. The effect, for the developers and QA team who share a space with 14 workstations, is that neither public nor private workstyles are facilitated. Noise and distraction are significant problems, as is privacy. Conversations by co-workers, phones (particularly cell phones), pagers, etc., are constant. At the same time, it's impossible to carry on a private (or discrete) phone call at my own desk.
Low walls essentially turn the space into an open-plan office, with various group and individual workstyles meaning that there are radically different work patterns. Neither privacy (a quite, distraction-free personal work environment) nor a small-group bull-pen are possible. Other departments have similar or worse issues with major corridors walking through workspaces, mismatches of proximity and projects -- work teams being split up or disparate teams being pushed into the same space.
It's a less than optimal situation, and working from home (lots of windows, radio, large desk, network connectivity, remote access, phone, IRC) make it an attractive alternative. We do have a large number of conference rooms, and the space is still being built out, so there's hope for improvement.
This is lifted from Gregory Pfister's In Search of Clusters: the ongoing battle in lowly parallel computing. Among the discussions: caching systems work best by randomly clearing sections in the cache (desktop), which should remain fully populated. Emptying the cache (clearing your desktop) wastes cycles. In the section "The Cache as a Messy Desk" (6.2.7, p 146):
Suppose you walk into an office and start to work on some project. You dig the files out of a desk drawer or cabinet, put them on your desk, and work. Then you start something else. No, don't put the old stuff away; restrain your tidiness even if you completely finished the previous job. just get out the new things you need, find an empty spot on your desk, and works there....Eventually, of course, your desk will be amess with no more room left on it. That's when youclear some space off by putting some of the itmes on your desk back in the files -- but only clear enough space to do the next theing you need.
Notice that the desk never, ever gets clean. In fact, except for the initial period when you first start using it, the desk is always completely filled with "old stuff" of various ages....
Cache memories work exactly like that messy desk; Noting is emptied out of them until the space it occupies is required for something else.
This is more of a legacy MS Windows issue than a Linux one. It also does not apply to flatscreen (LCD) panel displays.
The default refresh rate on many CRTs is 60Hz. This creates an interference pattern with any proximate florescent lighting, and is the major cause of "monitor glare", which isn't in fact glare, but flicker. It's particularly noticable in peripheral vision -- I often find I'm more disturbed by a co-worker's display (caught from the corner of my eye) than they appear to be.
The solution is to set monitor refresh to some higher level -- 72Hz or better seems to push the effect beyond any perceptible threshhold.
In Linux, check the output of your X server (startx -- 1>.startx.log 2>&1 & should save output to '.startx.log) to see what rates are being selected for your X session, or monitor the rates with tools like xvidtune or your window manager or desktop video tuning tools.
I'll have to read the judge's ruling and the statute, but my understanding is that an injunction is backed by a bond (or perhaps other terminology, but same effect) placed by the plaintif, in proportion to damages. I suspect a similar arrangement in this case. Statute would be 17 USC 502, though I don't see reference to the bond here. Hmmm....
IANAL. Corrections/amplifications apprecitated.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Of course, with/., we don't know what goes on in the sub queue, or if/when it's being attacked. It would be interesting to know about this, though there is the copycat problem associated with asking the question.
And I've got to say,/.'s been a great friend of K5 today, shout out to VA as well. Thanks, people. The world may not be perfect, but parts of it are excellent.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Rusty pulled K5 "because I didn't want my name associated with what was showing up on the site". The issue was discriminating signal from noise. At a certain point, things reached the level of crashing scoop.k5.org, but this wasn't the initial or principle problem.
The problem is that IP-based blocking only works against finite IPs. In this case, the attacks were coming from a relatively small number of sites, but things kept escalating beyond the ability of the K5 volunteer staff to deal with them.
Yes, chokepoint DoS is a possible attack, but the weblog was choking on poor quality data long before that.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
What pains me is that the Three Big Weblogs (TBW) have portions of the solution. Slashdot has filtering tools. K5 has a good moderation system. Advogato has a good membership vetting system. However, the pieces need to be put together. Having them on seperate systems doesn't quite cut it.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
I posted a rough set of notes on what I felt are the components of a good moderation system at scoop (http://scoop.kuro5hin.org/). Unfortunately, the site's down at the moment.
Abstracting, the Scoop engine uses a bounded metric (floating point 1-5 score) plus editorial oversite (content can be removed) to filter content.
Some of the interface tools need to be improved. Bulk moderation (set scores, then submit en mass) and filtering (seting min/max thresholds) need to be implemented. There's also the whole issue of anonymous story and content submission -- I ultimately feel that a solution akin to that described by Larry Lessig in Code and Other Laws of Cyberspace, the "Yale Wall", is necessary. This describes a physical posting board on which anonymous posts were allowed (not garbage-collected), if signed, potentially by anyone. Weblog equivalent would be an anon queue, regularly cleaned out, in which registered users could "sign" posts, but wouldn't be obligated to. Anonimity is then a grant by the community, but isn't a fully free of responsibility.
I do feel somewhat strongly that there has to be an equivalent of what's called "karma" at/., though the past reaction has been rather strongly negative when the issue's been raised at K5. Advogato's trust metric is one implmentation, I think it's better than/.'s, but I haven't seen something that works really well yet.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
I'd specifically excluded all of/., including its ad server, from my junkbuster filters. Sorry, but distracting animated banners just don't fly (and this is with the gif animation toggle-munged version of Netscape -- sufficiently long animation cycles will appear, though they won't loop. I've reluctantly added/.'s ad server to my blockfile this weekend.
This is unfortunate, because it's actually useful for me to keep tabs on who's advertising what at Slashdot.
I've already written Rob & Jeff suggesting they review acceptable copy guidelines. Any animation cycling < ~1 second is seriously distracting. More slowly cycling gifs aren't nearly as bad. My skin still crawls remembering a banner which ran almost a year ago with some guy's eyes strobing out of his head (cartoon-style) looking at a bill, advertising low service rates for some luser company or another. That directly prompted my adoption of Junkbuster and other means of regaining control of my browsing.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
The other value of multiple windows is that you can choose different display settings for, say, full page vs. a thread.
For full page, I tend to set threshold at 2 (or 3, if things are really bad), mode to threaded, highest moderated and most recent first. For a thread, I prefer threshold=0 (filter nothing but trolls), and mode to nested. I do this by hand-editing the URL (eg: http://slashdot.org/comments.pl?sid=00/07/04/04302 49&mode=nested&threshold=0&cid=46), and changing the cid to view different threads.
It would be kinda nice to be able to set seperate front page and thread view modes;-)
What part of "Gestalt" don't you understand? Scope out Kuro5hin
The problem with w3m is that it has to load the page fully in order to know how to render it. Somewhat like Netscape. You buy some time by skipping graphics, but lose otherwise. I think you're stuck.
Mozilla (and IE 5.x) both render and redraw on the fly. This makes for much faster percieved page loads and redraws (say if you resize a window). If the console-mode browsers could emulate this behavior, things might be cool, but I'm not sure how 'zactly you'd go about doing that.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
There is a place for regional shows of Linux, technology, firms, and benefits. These should be geared to the level of local interest, however. We don't need every town of 100,000 with a LUG trying to sponsor LinuxWorld Expo. Good shows are a lot of work for all involved, including the organizers, vendors, keynotes, hands-on instructors, and attendees.
One model which seems appropriate to me is something akin to an installfest on steroids, aimed at low- and mid-tier local companies, showing what Linux is, what it does, how it can be installed, and how it can solve business needs. There have been several of these, aimed at gathering a few hundred local businesses plus local Linux talent. The forum is to showcase local Linux firms and consultants. There's also the option to bring in a few national firms for a low-key presence - a small demo but not a full-booth setup, with plenty of opportunities to interact with local businesses directly.
Linux is about fitting to the tool to the task -- let's think appropriate technology, or forums, here.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
This is a Good Book (TM), and well worth reading, and timothy's writeup is solid. However, I was turned off initially by what I still consider to be an inexcusable failing of the book.
Per my standard practice, after cracking the spine, I went to first the ToC, then the back of the book -- the index. Entries for Microsoft, Apple, Xerox, PARC, IBM, Intel, Sun, Bill Gates, Steve Jobs, etc., etc., etc.
Entries for: Linux, Free Software Foundation, GNU, Linus Torvalds, Richard Stallman, Eric S. Raymond, Open Source, Apache -- nil. Ponder this: how many pages are printed worldwide by Xerox copiers in a week? How many pages are served worldwide by Apache webservers in an hour? "The document company" is completely dissing the Internet -- the largest, most accessible, and most efficient document distribution system ever invented?
The fact that a book could be published in the year 2000 with no references to the largest sea-change to sweep computing and IT in two decades, well into its mainstream adoption curve, is mind boggling. I'm not sure whether it's a failing of the indexers (though I don't recall specific mentions of any free software technologies, though the 'Web is referenced once or twice), the authors, or simply an example of failed vision at PARC. I remain simply stunned.
That said, where the book does go, it's good. By and large, it's an argument for many of the dynamics which make free software work. FS is a social invention as much as a technical one, and as much as our interfacing occurs over the web, email, and (sometimes) phone, I've also met some good friends FTF at local LUGs, regional meetings, and on occasions when paths crossed, even when oceans were bounded in the process.
Read this book, but read it critically.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Activities other than copying, distribution and modification are not covered by this License; they are outside its scope.
The act of running the Program is not restricted, and the output from the Program is covered only if its contents constitute a work based on the Program (independent of having been made by running the Program). Whether that is true depends on what the Program does.
(Emphasis added).
I would argue that the GPL doesn't explicitly grant public display and performance rights, but implicitly grants these rights through the highlighted language above.
A question is whether or not the GPL grants rights unless otherwise denied (subject to compliance requirements), or only grants specific rights. My reading of the license, including the Section 0: language above, is that all rights required to run a program are granted unless otherwise restricted, and that only copying, distribution, and modification are specifically restricted.
One of the issues which many would like to have addressed in GPL v3 is a linkage of distribution triggers to web-based, embedded, and CORBA-linked applications. These constitute a significantly different copyright environment than the one in which the GNU GPL was originally written.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Quibbles on Ben's analysis, areas of agreement ignored:
Term 1: a nonexclusive grant of copyright is conferred. In order to make copies, distribute them, and/or modify them, a copy right must be granted. However, it's nonexclusive -- that is, the grant doesn't transfer rights from holder to designee, but extends them in a limited sense.
Term 4: additional terms. If the original grant is under additional terms (or multiple terms), essentially dual licensing as is practiced with Perl or allowed under MozPL, then the GNU GPL doesn't provide for transmission of the alternative terms, but the other license may. Dual licensing may be a useful tool for free software distribution and compatibility licensing between different codebases.
Terms 11 & 12. More significant than perhaps Mr. Tilly may appreciate. Think third-party claims under liability or warrantee. Does the GPL put such a third-party claimee into the position of being able to press suit or be in compliance with copyright? Hmmm.... I'm not sure, but I've got to talk to my lawyer.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
While you could publish your own edition of AAIW, I'm not sure you could list yourself as the author, in the US, under attribution obligations of copyright. You certainly couldn't claim copyright protection for your own verbatim reproduction.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Alan Knowles, an independent developer, is working on a PHP port of our OpenMerchant project. You can find information on OpenMerchant-PHP at http://www.hklc.com/projects/.
Weblogs are a cool concept, but ultimately lead to fragmentation -- content, eyeballs, authors, and participants are spread among many distinct islands.
One of the more interesting ideas to emerge from the Advogato / Kuro5hin axis is the concept of syndication. This would cover content, already common -- Slashdot and LinuxToday are essentially content syndication sites, and The Register officially sanctions linking. But syndication could also include a distributed user directory, and potentially (flame on) attributes such as karma or other metrics of merit from various sites.
I see a mix of several models coallescing into the final "product":
From Slashdot and Blockstackers -- Everything -- a hyperlinked, persistant, discussion/directory. Somewhat like Wiki.
From Advogato, the idea of a trust metric is useful, but not sufficient.
From the IWETHEY EZBoard, active content promotion. Active topics float up in the discussion queue. It's a bit different from a typical weblog, but tends to promote issues of interest and bury (but not kill) those which aren't generating much traffic.
Still to be worked out are issues of story selection. Various models work -- Slashdot and IWETHEY fall at two extremes, with a dedicated editorial staff on the one hand, and a number of free-form "open forums" in which any topic may be posted and discussed. Kuro5hin's still working out the kinks, though a number of suggestions have been proposed.
The point is that high-quality (and low quality) content are created all over the Net. Mindless Link Propogation (TM) (MLP) is a useful way of aggregating it to key sites. Mindful link propogation might be even better.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Eben Moglen, Columbia Law School professor and pro bono lawyer for the FSF, spoke on MarketPlace Radio (Minnesota Public Radio, carried on many NPR stations) last Wednesday following the court's breakup finding. One of Moglen's comments was that the FoF against Microsoft, detailing as it did a list of anti-competitive and monopolistic activities, would preclude Microsoft from being able to assert its patent IP rights.
I don't recall the precise argument, and MarketPlace doesn't provide freely accessible archives, but it might be that Microsoft would find itself stripped of its patents or simply unable to assert them. Patents are a legal means for providing monopoly advantage in a restricted area for a limited time -- an advantage courts might find the company doesn't require or has lost rights to claim.
This does raise the bar though for any release of source or APIs to Microsoft code, however. Interesting times, it seems.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Mounting is built in to most Linux desktop environments. Works pretty much like the Macintosh, BTW, which is geared to ease of use, and, also BTW, is based on the basic premise of "mounting" removable media.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
The significance of a trade or certification mark is control over branding. Forking the code, extending it, whatever, is allowed, but, as an example, Microsoft cannot call its forked Kerberos "Kerberos", if the trademark-holding authority says that they can't (note that a trademark search at the USPTO's website for "kerberos" produces no hits).
The trick is that the trademark owner can establish his/her own standards for licensing, which can be loose (Linux (TM)), strict (Xerox (TM)), or subject to meeting some standard (Harris Tweed (TM), Underwriters Laboratories (TM)). And if the mark holder wants to change the terms (to head off noncompliant extensions), it's their right to do so.
Licensing terms are an arbitrary decision on the part of the mark holder. The obligation of the markholder, however, is to uphold the mark. Trademarks can be lost -- asprin is one, dixie cups (IIRC) is another. A lawyer friend once had to do research in defending the Hooters (TM) trademark (I kid you not).
What we've seen in the past are modifications to code which wasn't subject to trademark (kerberos), or attempts to regulate ability to modify code directly, rather than certification of compliance (Java). Neither mode works particularly well, as we've seen.
The use of a mark to insure compliance means that someone contemplating a code fork has to weigh the strategic advantages of noncompliant operation with the loss of branding or certified compliance. Likewise, the licensing authority is under pressure to keep terms reasonable enough that a seperate compliance program isn't launched in competition, with more reasonable (or easier to comply with) rules.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Excellent point, and it shows the importance of trademark and certification mark in free software. While there's a strong argument to distributing spec and code in free software, there's an equally strong argument to retaining and reserving the rights to the trade name of the product being released.
One of the better ways to do this is to provide a regression test for the software in question, requiring it to meet the test. Microsoft, in this world, would be free to embrace and extend all it wanted, and would have free access to the source to do this. However, unless they met the terms for the trade name usage, they would have to call the product by some other name -- defeating the whole marketroid check-off item school of product promotion.
Many free software firms and propenents don't yet see the importance of this. Sun Microsystems in particular has just plain got it wrong in attempting to enforce compliance through code regulation rather than certification marks and compliance/compatibility testing. Very frustrating, as I've worked with some of the folks involved in the SCSL concept. IMO most of the issues Sun has seen in licensing Java would be non-starters had they used an alternative approach. As things stand, they're fighting with Microsoft for the standard, while IBM emerges as the market leader in workable Java implementations.
What part of "Gestalt" don't you understand? Scope out Kuro5hin
Slashdot Mirror
...in about two weeks. Looking for a Sept. relaunch.
But thanks for paying attention.
What part of "Gestalt" don't you understand?
Aside from some dated information and inaccuracies, the problem I've had with the LMD for quite some time is that it's couched as a double negative: "Linux isn't evil foo", where foo is some undesirable characteristic. The entire flavor of the document would change if language were changed from negative to neutral or even positive. Otherwise it has this "are you still beating your wife?" flavor. The LMD would answer that question with a topic "Linux is no longer beating its wife".
Just as an example, the document would have an entirely different flavor if the system headings under "4. Systems Myths" were:
Suggestion is a powerful tool. The document should suggest that Linux is the cat's pyjamas. People will tend to believe.
What part of "Gestalt" don't you understand?
Interesting to see your post, I've got my own copy of the book open to Section II, dealing with the issue of workspaces.
There are those who've always held that Peopleware (Tom DeMarco & Timothy Lister, Dorset House Publishing Co., New York, © 1987 ISBN: 0-932633-05-6) is a programmer's wet dream of the ultimate work environment. I'd also say that things have changed a lot in the 13 years since it was published: the Internet, a workforce which has grown up in cube farms, voicemail and pagers replacing overhead paging. Still, there are some powerful truths, argued with facts, contained in the book, and echoed in more recent works such as Steve McConnell's Code Complete.
I do strongly feel that open-plan and private-office workspaces fall at either end of a large continuum, and probably are not ideal for much contemporary technical work. The shared office or high-walled cube approach seems to work better than either open or private layouts.
Of the environments I've worked in, two of the most conducive where a "pod" arrangement of cubes, in which four cubes shared a common corridor access. Scooting back on my chair I could convene with co-workers on similar projects, yet within my cube, I had relative privacy. The other was a "coding pen" with three developers at standard desks pushed up against walls. Small enough group that multi-way communications weren't overly common, enough people that collaboration and quick queries were possible.
One of the worst environments was at Visa International, which has "rationalized" its contractor workspace environment by placing all contractors on a single floor of its Foster City offices, dubbed the UN for its international flavor. My direct reports were six floors above me, and co-worker contractors were a dozen cubes down the aisle and across the floor. Hardly conducive to communications (but I got the skinny on the Y2K project and the marketing design campaign).
Private offices, in a development enviroment, tend to cut too many communications channels. The worst situation is closed doors and no glass -- it's hard to tell what you're walking into and/or whether or not it's appropriate. Even a cube-orientation can be almost as effective. The same office which originally had a "coders pen" moved to a new location in which a high-walled four-workstation unit shaped like a "+" was the initial furnishings. Interactions were virtually stifled: the only ways to communicate were to walk around the entire periphery, or to shout in a voice audible over the entire office. Later, a section of cubes with mixed wall heights was adopted -- communications between common working groups was possible, distractions from other groups were minimized.
With the Internet, the opportunities for distraction have also increased. As you say, downloads, pr0n, computer games, personal business, are all distractions. I don't feel that this is the best mode of operation for most technical work, particularly for people who aren't used to working in such an autonomous environment, unless a well defined (and resonably revisable) set of performance milestones can be established, in which case I really don't care where you work or how you spend your time, so long as you produce.
My current employer has just moved from a large-office, high-walled cube environment to a workspace typified by smaller office spaces (8 to 20 workstations), with low cube walls -- about 5'. The effect, for the developers and QA team who share a space with 14 workstations, is that neither public nor private workstyles are facilitated. Noise and distraction are significant problems, as is privacy. Conversations by co-workers, phones (particularly cell phones), pagers, etc., are constant. At the same time, it's impossible to carry on a private (or discrete) phone call at my own desk.
Low walls essentially turn the space into an open-plan office, with various group and individual workstyles meaning that there are radically different work patterns. Neither privacy (a quite, distraction-free personal work environment) nor a small-group bull-pen are possible. Other departments have similar or worse issues with major corridors walking through workspaces, mismatches of proximity and projects -- work teams being split up or disparate teams being pushed into the same space.
It's a less than optimal situation, and working from home (lots of windows, radio, large desk, network connectivity, remote access, phone, IRC) make it an attractive alternative. We do have a large number of conference rooms, and the space is still being built out, so there's hope for improvement.
What part of "Gestalt" don't you understand?
This is lifted from Gregory Pfister's In Search of Clusters: the ongoing battle in lowly parallel computing. Among the discussions: caching systems work best by randomly clearing sections in the cache (desktop), which should remain fully populated. Emptying the cache (clearing your desktop) wastes cycles. In the section "The Cache as a Messy Desk" (6.2.7, p 146):
Suppose you walk into an office and start to work on some project. You dig the files out of a desk drawer or cabinet, put them on your desk, and work. Then you start something else. No, don't put the old stuff away; restrain your tidiness even if you completely finished the previous job. just get out the new things you need, find an empty spot on your desk, and works there....Eventually, of course, your desk will be amess with no more room left on it. That's when youclear some space off by putting some of the itmes on your desk back in the files -- but only clear enough space to do the next theing you need.
Notice that the desk never, ever gets clean. In fact, except for the initial period when you first start using it, the desk is always completely filled with "old stuff" of various ages....
Cache memories work exactly like that messy desk; Noting is emptied out of them until the space it occupies is required for something else.
Viva la mess!
What part of "Gestalt" don't you understand?
This is more of a legacy MS Windows issue than a Linux one. It also does not apply to flatscreen (LCD) panel displays.
The default refresh rate on many CRTs is 60Hz. This creates an interference pattern with any proximate florescent lighting, and is the major cause of "monitor glare", which isn't in fact glare, but flicker. It's particularly noticable in peripheral vision -- I often find I'm more disturbed by a co-worker's display (caught from the corner of my eye) than they appear to be.
The solution is to set monitor refresh to some higher level -- 72Hz or better seems to push the effect beyond any perceptible threshhold.
In Linux, check the output of your X server (startx -- 1>.startx.log 2>&1 & should save output to '.startx.log) to see what rates are being selected for your X session, or monitor the rates with tools like xvidtune or your window manager or desktop video tuning tools.
What part of "Gestalt" don't you understand?
I'll have to read the judge's ruling and the statute, but my understanding is that an injunction is backed by a bond (or perhaps other terminology, but same effect) placed by the plaintif, in proportion to damages. I suspect a similar arrangement in this case. Statute would be 17 USC 502, though I don't see reference to the bond here. Hmmm....
IANAL. Corrections/amplifications apprecitated.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Of course, with /., we don't know what goes on in the sub queue, or if/when it's being attacked. It would be interesting to know about this, though there is the copycat problem associated with asking the question.
And I've got to say, /.'s been a great friend of K5 today, shout out to VA as well. Thanks, people. The world may not be perfect, but parts of it are excellent.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Rusty pulled K5 "because I didn't want my name associated with what was showing up on the site". The issue was discriminating signal from noise. At a certain point, things reached the level of crashing scoop.k5.org, but this wasn't the initial or principle problem.
The problem is that IP-based blocking only works against finite IPs. In this case, the attacks were coming from a relatively small number of sites, but things kept escalating beyond the ability of the K5 volunteer staff to deal with them.
Yes, chokepoint DoS is a possible attack, but the weblog was choking on poor quality data long before that.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
What pains me is that the Three Big Weblogs (TBW) have portions of the solution. Slashdot has filtering tools. K5 has a good moderation system. Advogato has a good membership vetting system. However, the pieces need to be put together. Having them on seperate systems doesn't quite cut it.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
I posted a rough set of notes on what I felt are the components of a good moderation system at scoop (http://scoop.kuro5hin.org/). Unfortunately, the site's down at the moment.
Abstracting, the Scoop engine uses a bounded metric (floating point 1-5 score) plus editorial oversite (content can be removed) to filter content.
Some of the interface tools need to be improved. Bulk moderation (set scores, then submit en mass) and filtering (seting min/max thresholds) need to be implemented. There's also the whole issue of anonymous story and content submission -- I ultimately feel that a solution akin to that described by Larry Lessig in Code and Other Laws of Cyberspace, the "Yale Wall", is necessary. This describes a physical posting board on which anonymous posts were allowed (not garbage-collected), if signed, potentially by anyone. Weblog equivalent would be an anon queue, regularly cleaned out, in which registered users could "sign" posts, but wouldn't be obligated to. Anonimity is then a grant by the community, but isn't a fully free of responsibility.
I do feel somewhat strongly that there has to be an equivalent of what's called "karma" at /., though the past reaction has been rather strongly negative when the issue's been raised at K5. Advogato's trust metric is one implmentation, I think it's better than /.'s, but I haven't seen something that works really well yet.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Persistant buggers out there.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
I'd specifically excluded all of /., including its ad server, from my junkbuster filters. Sorry, but distracting animated banners just don't fly (and this is with the gif animation toggle-munged version of Netscape -- sufficiently long animation cycles will appear, though they won't loop. I've reluctantly added /.'s ad server to my blockfile this weekend.
This is unfortunate, because it's actually useful for me to keep tabs on who's advertising what at Slashdot.
I've already written Rob & Jeff suggesting they review acceptable copy guidelines. Any animation cycling < ~1 second is seriously distracting. More slowly cycling gifs aren't nearly as bad. My skin still crawls remembering a banner which ran almost a year ago with some guy's eyes strobing out of his head (cartoon-style) looking at a bill, advertising low service rates for some luser company or another. That directly prompted my adoption of Junkbuster and other means of regaining control of my browsing.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
The other value of multiple windows is that you can choose different display settings for, say, full page vs. a thread.
For full page, I tend to set threshold at 2 (or 3, if things are really bad), mode to threaded, highest moderated and most recent first. For a thread, I prefer threshold=0 (filter nothing but trolls), and mode to nested. I do this by hand-editing the URL (eg: http://slashdot.org/comments.pl?sid=00/07/04/04302 49&mode=nested&threshold=0&cid=46), and changing the cid to view different threads.
It would be kinda nice to be able to set seperate front page and thread view modes ;-)
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
The problem with w3m is that it has to load the page fully in order to know how to render it. Somewhat like Netscape. You buy some time by skipping graphics, but lose otherwise. I think you're stuck.
Mozilla (and IE 5.x) both render and redraw on the fly. This makes for much faster percieved page loads and redraws (say if you resize a window). If the console-mode browsers could emulate this behavior, things might be cool, but I'm not sure how 'zactly you'd go about doing that.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
There is a place for regional shows of Linux, technology, firms, and benefits. These should be geared to the level of local interest, however. We don't need every town of 100,000 with a LUG trying to sponsor LinuxWorld Expo. Good shows are a lot of work for all involved, including the organizers, vendors, keynotes, hands-on instructors, and attendees.
One model which seems appropriate to me is something akin to an installfest on steroids, aimed at low- and mid-tier local companies, showing what Linux is, what it does, how it can be installed, and how it can solve business needs. There have been several of these, aimed at gathering a few hundred local businesses plus local Linux talent. The forum is to showcase local Linux firms and consultants. There's also the option to bring in a few national firms for a low-key presence - a small demo but not a full-booth setup, with plenty of opportunities to interact with local businesses directly.
Linux is about fitting to the tool to the task -- let's think appropriate technology, or forums, here.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
This is a Good Book (TM), and well worth reading, and timothy 's writeup is solid. However, I was turned off initially by what I still consider to be an inexcusable failing of the book.
Per my standard practice, after cracking the spine, I went to first the ToC, then the back of the book -- the index. Entries for Microsoft, Apple, Xerox, PARC, IBM, Intel, Sun, Bill Gates, Steve Jobs, etc., etc., etc.
Entries for: Linux, Free Software Foundation, GNU, Linus Torvalds, Richard Stallman, Eric S. Raymond, Open Source, Apache -- nil. Ponder this: how many pages are printed worldwide by Xerox copiers in a week? How many pages are served worldwide by Apache webservers in an hour? "The document company" is completely dissing the Internet -- the largest, most accessible, and most efficient document distribution system ever invented?
The fact that a book could be published in the year 2000 with no references to the largest sea-change to sweep computing and IT in two decades, well into its mainstream adoption curve, is mind boggling. I'm not sure whether it's a failing of the indexers (though I don't recall specific mentions of any free software technologies, though the 'Web is referenced once or twice), the authors, or simply an example of failed vision at PARC. I remain simply stunned.
That said, where the book does go, it's good. By and large, it's an argument for many of the dynamics which make free software work. FS is a social invention as much as a technical one, and as much as our interfacing occurs over the web, email, and (sometimes) phone, I've also met some good friends FTF at local LUGs, regional meetings, and on occasions when paths crossed, even when oceans were bounded in the process.
Read this book, but read it critically.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Section 0 of the GNU GPL states:
(Emphasis added).
I would argue that the GPL doesn't explicitly grant public display and performance rights, but implicitly grants these rights through the highlighted language above.
A question is whether or not the GPL grants rights unless otherwise denied (subject to compliance requirements), or only grants specific rights. My reading of the license, including the Section 0: language above, is that all rights required to run a program are granted unless otherwise restricted, and that only copying, distribution, and modification are specifically restricted.
One of the issues which many would like to have addressed in GPL v3 is a linkage of distribution triggers to web-based, embedded, and CORBA-linked applications. These constitute a significantly different copyright environment than the one in which the GNU GPL was originally written.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
IANAL also.
Quibbles on Ben's analysis, areas of agreement ignored:
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
While you could publish your own edition of AAIW, I'm not sure you could list yourself as the author, in the US, under attribution obligations of copyright. You certainly couldn't claim copyright protection for your own verbatim reproduction.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Alan Knowles, an independent developer, is working on a PHP port of our OpenMerchant project. You can find information on OpenMerchant-PHP at http://www.hklc.com/projects/.
Me -- I work for OpenSales.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Weblogs are a cool concept, but ultimately lead to fragmentation -- content, eyeballs, authors, and participants are spread among many distinct islands.
One of the more interesting ideas to emerge from the Advogato / Kuro5hin axis is the concept of syndication. This would cover content, already common -- Slashdot and LinuxToday are essentially content syndication sites, and The Register officially sanctions linking. But syndication could also include a distributed user directory, and potentially (flame on) attributes such as karma or other metrics of merit from various sites.
I see a mix of several models coallescing into the final "product":
Still to be worked out are issues of story selection. Various models work -- Slashdot and IWETHEY fall at two extremes, with a dedicated editorial staff on the one hand, and a number of free-form "open forums" in which any topic may be posted and discussed. Kuro5hin's still working out the kinks, though a number of suggestions have been proposed.
The point is that high-quality (and low quality) content are created all over the Net. Mindless Link Propogation (TM) (MLP) is a useful way of aggregating it to key sites. Mindful link propogation might be even better.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Eben Moglen, Columbia Law School professor and pro bono lawyer for the FSF, spoke on MarketPlace Radio (Minnesota Public Radio, carried on many NPR stations) last Wednesday following the court's breakup finding. One of Moglen's comments was that the FoF against Microsoft, detailing as it did a list of anti-competitive and monopolistic activities, would preclude Microsoft from being able to assert its patent IP rights.
I don't recall the precise argument, and MarketPlace doesn't provide freely accessible archives, but it might be that Microsoft would find itself stripped of its patents or simply unable to assert them. Patents are a legal means for providing monopoly advantage in a restricted area for a limited time -- an advantage courts might find the company doesn't require or has lost rights to claim.
This does raise the bar though for any release of source or APIs to Microsoft code, however. Interesting times, it seems.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Mounting is built in to most Linux desktop environments. Works pretty much like the Macintosh, BTW, which is geared to ease of use, and, also BTW, is based on the basic premise of "mounting" removable media.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
The significance of a trade or certification mark is control over branding. Forking the code, extending it, whatever, is allowed, but, as an example, Microsoft cannot call its forked Kerberos "Kerberos", if the trademark-holding authority says that they can't (note that a trademark search at the USPTO's website for "kerberos" produces no hits).
The trick is that the trademark owner can establish his/her own standards for licensing, which can be loose (Linux (TM)), strict (Xerox (TM)), or subject to meeting some standard (Harris Tweed (TM), Underwriters Laboratories (TM)). And if the mark holder wants to change the terms (to head off noncompliant extensions), it's their right to do so.
Licensing terms are an arbitrary decision on the part of the mark holder. The obligation of the markholder, however, is to uphold the mark. Trademarks can be lost -- asprin is one, dixie cups (IIRC) is another. A lawyer friend once had to do research in defending the Hooters (TM) trademark (I kid you not).
What we've seen in the past are modifications to code which wasn't subject to trademark (kerberos), or attempts to regulate ability to modify code directly, rather than certification of compliance (Java). Neither mode works particularly well, as we've seen.
The use of a mark to insure compliance means that someone contemplating a code fork has to weigh the strategic advantages of noncompliant operation with the loss of branding or certified compliance. Likewise, the licensing authority is under pressure to keep terms reasonable enough that a seperate compliance program isn't launched in competition, with more reasonable (or easier to comply with) rules.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin
Excellent point, and it shows the importance of trademark and certification mark in free software. While there's a strong argument to distributing spec and code in free software, there's an equally strong argument to retaining and reserving the rights to the trade name of the product being released.
One of the better ways to do this is to provide a regression test for the software in question, requiring it to meet the test. Microsoft, in this world, would be free to embrace and extend all it wanted, and would have free access to the source to do this. However, unless they met the terms for the trade name usage, they would have to call the product by some other name -- defeating the whole marketroid check-off item school of product promotion.
Many free software firms and propenents don't yet see the importance of this. Sun Microsystems in particular has just plain got it wrong in attempting to enforce compliance through code regulation rather than certification marks and compliance/compatibility testing. Very frustrating, as I've worked with some of the folks involved in the SCSL concept. IMO most of the issues Sun has seen in licensing Java would be non-starters had they used an alternative approach. As things stand, they're fighting with Microsoft for the standard, while IBM emerges as the market leader in workable Java implementations.
What part of "Gestalt" don't you understand?
Scope out Kuro5hin