First, I would not call subcontracting to a reasonably decent design bureau in one of the two most advanced space technology powers outsourcing.
Second, as far as cost and POC design Russians are a better choice the Americans. They generally tend to do loads of POC work instead of at-desk design and modelling (just look how many different POCs were done for the Buran for example). As a result they are much better at understanding the concept of a POC and doing it cheap and cheerful without unnecessary overengineering. If it has to fly for real first time and money is (nearly) unlimited it makes sense to hire Boeing or someone similar. But not for a POC. Russians do it better. It is the way they do stuff.
People keep forgetting why liquid metal cooling is being abandoned in the nuclear industry. Liquid metals tend to be extremely aggressive substances. A Gallium-Indium mix will dissolve nearly any metal or alloy over time. Ceramics and glass tend to get permeated and lose their mechanical properties. Frankly no idea about plastics.
So, let's play the devil advocate here: how does this differ from the market for compiler tools on Linux? I am giving this example because once upon a time there was such a lawsuit by some lousy lamers with a pascal compiler against all major unix vendors. As a result they stopped shipping C without extra licenses. Somewhere in the late 80-es I believe.
The effort into trying to cure their "deviations" has been minimal. In fact none.
There will be no real effort into curing their "deviations" for a long time to come because any attempt to do serious scientific research into it will get torpedoed by gay rights groups. Can't blame 'em. A cure for the urge to shag a 5 year old boy may end up being a cure for the urge of shagging any person of the same sex. Some of them do not like the idea of this cure being available because it is likely that it will be forced on at least some people. At least that is the official reason. The real reason is that there is no way to collect enough experimental data from sex offender cases while other "deviations"...
The only moderately successfull treatment has been stuffing them with chemicals to suppress any sexual urge. Scandinavian countries have been doing this for 10+ years now and have a nearly zero rate of pedofiliac associated crime (abduct, shag and kill cases) and nearly zero repeat offence rate. Other countries (France) are looking into it. Unfortunately neither UK, nor US have been even considering this. The obvious problem with this approach is that the moment the drugs stop the person is likely to reoffend.
So at least until we start really looking into the possibility of altering sexual behaviour via therapy the sex offenders are here to stay. If we do not lock them and do not stuff them with chemicals we might as well tag them. Not that it will help. It was considered in Sweden and France as an alternative to stuffing them with sexual depressants and it was found that when under the urge they do not give a flying fuck about any control measures. They will just do it, so the right way of dealing with it is to ensure they never get the urge in first place.
Reuters does not put all of its stuff on the web. It is a mixed subscription + free model.
Sun is not a newspaper. It is a form of porn disguised as news. It does not serve the same audience.
The Mirror is not a newspaper. It is a form of prudish porn disguised as news. It also does not server audience that plans to read news.
Times is planning to go to a subscription model. In fact, I think they are trialing a system for charging non-UK subscribers.
This in fact leaves only the Guardian and the Scotsman who are doing guess what? Campaigning for the extermination of news.bbc.co.uk as unfair competition.
There is a long and ongoing conflict between the BBC and the other online media in the UK. BBC has been able to nearly exterminate all UK online news except the specialized ones. Even those have nowdays moved to a subscription model (FT, DT and the like). It is free, it is high quality and it is updated round the clock. This is seriously annoying most newspaper owners and editors who see their circulation dropping and online presence being pushed into the oblivion. Allowing them to feed on an RSS feed is a great marketing ploy. They instead of protesting endlessly will now become largely dependant on BBC news online existing. Once this happens Blair (or any other creature which replaces him) will no longer have the option to exterminate news.bbc.co.uk for "efficiency of public spending" reasons.
Well... Most S.M.A.R.T. temperature sensors are on the PCB and they are measuring PCB temperature instead of the internal drive temperature.
Hence, a fan under the disk makes a lot of difference while making very little to make your data safer.
A 3x 40mm fan battery in front of a drive or a pressed enclosure that cools the actual package holding the platters makes a lot of difference there while not chaning the S.M.A.R.T. reading by more then a degree or so.
It is up to you - what do you want. Show (a good reading) or substance (good temperature of your drive platters and heads).
Err... Making nice with Mandelson? I may be wrong, but I do not recall Allen being the specific orientation essential for servicing Mandelson's orifices.
There is also the bind8+ algorithm which is also followed by most other well behaved implementations. Besides the normal TTL criteria these will expire a record after every access with some non-zero probability. As a result records leave the cache before the TTL is reached.
This is done as a safeguard against typos in TTL and operational mistakes. Not everyone likes it. For example there was a long rant by DJB about why it sucks once upon a time. But it is good and nearly all popular implementations do it nowdays.
This has happened before. See for the discussions about LMcV and lmbench in the 90-es. In fact, the moment I saw that Linus has AGAIN selected a Larry McVoy tool my first thoughts were "Oh no, not another lmbench". I bet I was not the only one.
Considering that he is also known to be litigation happy I am not going to qualify his behaviour that time and this time. Just read the LKM on both occasions as well as some of his musings. They are selfexplanatory.
It was a mistake to agree to work with a company run by Larry McVoy in first place. Some of us remember previous McVoyisms from time past. Namely lmbench.
For a long time it was the best stress testing tool for Linux (and Unix for that matter). In fact if you look back around 1995 it was the de-facto standard for both performance and stress testing. That continued until Larry got pissed off at someone (forgot exactly what, look in lkm for the mid-nineties). So the tool went unmaintained and into oblivion.
There is no point to try working with people like this on any large project. They take stuff which is "just business" personally. They also get pissed off personally when someone finds a weakness in their way to do business and exploits it legally. So on so fourth.
My first reaction when BM surfaced 3 years ago was "Oh no... Not another lmbench". And I see that I was right...
That is correct. Another set of affected software which is not listed but affected is all opensource software using the old TAP/TUN driver. CIPE. OpenVPN 1.6, etc. All of these break. Interestingly enough some of them do not break immediately, but break sooner or later anyway.
1. Noisy - there is an environmental limit in the EU. It cannot be more than that. My car is fairly close to it and is tolerable on up to 5h trips so how to put it - do not really care.
2. Slow + short range - commuter car. Speed and range are not part of the spec. If it can move me the 5-10 miles to the office and back every day who cares.
Basically you are in the usual mistake of believing the top gear reviews. They are quite often valid, but suffer from the following serious deficiencies:
1. Car should fit Jeremy Clarkson in an American driving position. Despite being a "car guru" he still has not learned that you have to drive many cars in a french driving position (up and close to the wheel) to get the most of them. He drives splattered so he ends up not fitting in things like the Modus, Sirion, Copen, etc.
2. Car should require wrestling with it by Jeremy Clarkson to be driven. If it has a decent light electric power assisted steering it is blamed as lightweight, toylike and a few other adjectives.
3. So on so fourth.
Do not understand me wrong. They are still better then consumer reviews, but I would not trust them a lot. I would read carkeys or something a bit more balanced instead (where people who test know how to drive).
Show me a station offering free air at 300 bar. Wanna see one...
The basic problem with this car is that it will require extra infrastructure. Not terribly expensive, but quite noisy. Compressing air to 300 bar is not a very quiet affair.
They are getting better, but getting more narrow and more slanted towards a specific profession. For example the referred RFC is really funny only if you know some network engineering. If you do, you will end up in cramps on the floor from hysterical laughter. It is even better then RFC on IP over Avian Carriers with QoS and draft on Electicity over MPLS.
After all it is an IETF document - it is supposed to be funny to the IETF audience.
So what data does it have to write to disk if the bloody calendar is bloody empty? Similarly what data does it have to write if it is supposed to use a network calendar? So on so fourth.
It is a bug (and filed as such).
I buy from these guys: http://www.a1-battery.co.uk They are a front-end for several Chinese and Korean manufacturers.
If you are not in the UK get the part number from them and search in your country.
For example their Sony equivalents have 25%+ charge for nearly all models. I have not tried IBM and others but I would expect them to be the same. The only problem I have found is that they sometimes switch from trickle charge to full charge mode (actually I see it with all new batteries, my phone does the same). This results in annoying notifications "your battery is now full" once in a while.
I get 7+ hours out of an ancient Sony P3 1GHz bog standard desktop replacement class Vaio (3+ years old) under linux after throwing out the horrid puke sony ships for battery and replacing the battery and the CD with 2 x Chinese OEM 4800 mAh and using cpufreqd to keep the CPU frequency as low as possible when idling.
All I want is that the idiots at KDE HQ stop calling sync after each disk write operation to calendar and settings. I was almost ready to rebuild the entire thing with sync redefined as NOOP at one point. That will give me 1 more hour:-)
But definitely, do not see any case whatsoever for using an external battery. If you need that charge level simply throw out your CD and line up 1 or 2 cheap and chearfull Chinese or Korean OEM batteries. They quite often have 20-30% higher charge then the branded stuff and cost twice less.
The most interesting screenshot as far as potential users are concerned is missing. I do not see the screenshot of the mail client. While I am not a particular fan of IE, it works to some extent so the users of Win CE have a working browser.
What they do not have is a working mail client.
The built in portable outlook only fetches recent messages and only from Inbox. You cannot access IMAP folders other then INBOX and have no means of accessing your old mail. To add to that, Pocket PC Outlook on the Smartphone Edition does not have client side SSL certificate support for IMAP or POP3 so you cannot provide secure remote access. This is essential, because it does not support any (even lame) encrypted VPNs and there are no third party clients for it.
There are plenty of people out there who will _pay_ for a working thunderbird on Windoze Smartphones. And very few who actually care about a working third party browser to replace IE.
Well... as someone who has both of reviewed 3ware adapters in production I am not amazed. They are nice, but nothing to shout about. They also have LOADS of PROBLEMS not mentioned in the article.
8506 SATA series prior to a certain board revision are extremely susceptible to bus noise. As a result you have to find a way to bastardize the PCI bus down to 33MHz and provide additional grounding. Even so, they are likely to cause random system deaths and serious memory corruption in most Opteron MSI and Assus motherboards as well as some other designs. Using in 1U and 2U chassis with riser cards is a no-no for the same reason (exemption for some buffered risers). As a side note, most resellers will try to stuff you with an old board despite the fact that they know about this problem.
9506 board and linux driver at least as of 2.6.9 defaults to no read cache, only write cache which is outright daft. It is also the major reason for low performance at least under Linux.
Both are nice cards, but I would not recommend them to anyone who does not have extensive PC hardware knowledge. They are fussy, carpicious and very hard to troubleshoot when they go wrong.
Most SpaceShip one type efforts employ a subsonic plane "booster" because of the separation problem. There is no airframe design out there capable of performing all of the following: horizontal takeoff, lift, flight and successful separation at supersonic speeds while in the atmosphere. You either separate at subsonic speeds like the Pegasus and similar Russian efforts or you do the full monty and get out of the lower layers of the atmosphere and separate there the way all missile launchers do (first stage is usually to 20km at least).
If someone designs a separable airframe the technology to achieve Mach 3+ and 20km+ altitude is already there. Mig 31 in the reconnaissance version can do it, BlackBird can do it and these are 20 year old designs. While there is 22 Mach to go, if you look at the fuel expenditure doing 22 Mach from 20+ km altitude is likely to be about 10% of the fuel compared to doing 25 Mach from 0.
And anyway, methinks that spaceplane booster stages are the wrong idea. I think it will be cheaper to build a rail launcher somewhere at 4-5 km altitude that fires a missile payload at 10-20 degrees up. Accelerating a few hundred tons to 600km+ for a very short trip is no big deal by modern technical standards. There are places in the Andes, Rockies and Pamir that are reasonably flat, non-inhabited and have suitable roads and infrastructure close by at altitudes as high as 4km. It is only a matter of time until someone decides to do this.
It is not a personal bug. It is the showstopper bug as far as any network engineer or IT person is concerned.
There is no way to do a useable presentation with a network diagram in it as far as OO is concerned. Same for any more complex workflow, database, UML, you name it. They are all vector graphics which makes OO unusable in the industry most likely to be willing to look into adopting it (we all hate microshit).
OO uses for preview only raster out of EPS. Some EPS have an embedded low res raster for previews, some don't depends on the converter. Vector and PS proper are not visualised. So while EPS sometimes it works for printing copies you:
1. Lose WYSIYG. You do not see your diagram in relation to the text. Which is about the level of "userfriendliness" I get from LyX or even TeX/LaTeX. 2. Cannot use it for electronic viewing. Most importantly cannot do impress/powerpoint presentations. This is how I ran into this. I tried to do a network diagram which came out fairly reasonable in one of the recent Dia versions. After that I found out that there is no way in hell I can get that diagram into any powerpoint compatible presentation editor.
Well, you simply have not run into the other big OO gremlin yet.
Not a single vector graphics import format works properly.
The ones that barely work (Autocad for example) lose colors and most of the formatting. So if you want to draw a half decent diagram using DIA and import it into an OO presentation you might as well forget it. Your only chance is to export it as a raster image and import it in OO. The result is horrible by all means. Horrible size, horrible visually, horrible in a print form and horrible to edit.
And OO 2.0 does not fix a single one of this issues. Instead of that we get visual candy - KDE widget support. Excuse me, but can we actually get the basic functionality fixed first before we get into Clippy land.
Slashdot Mirror
First, I would not call subcontracting to a reasonably decent design bureau in one of the two most advanced space technology powers outsourcing.
Second, as far as cost and POC design Russians are a better choice the Americans. They generally tend to do loads of POC work instead of at-desk design and modelling (just look how many different POCs were done for the Buran for example). As a result they are much better at understanding the concept of a POC and doing it cheap and cheerful without unnecessary overengineering. If it has to fly for real first time and money is (nearly) unlimited it makes sense to hire Boeing or someone similar. But not for a POC. Russians do it better. It is the way they do stuff.
It will fail for other reasons.
People keep forgetting why liquid metal cooling is being abandoned in the nuclear industry. Liquid metals tend to be extremely aggressive substances. A Gallium-Indium mix will dissolve nearly any metal or alloy over time. Ceramics and glass tend to get permeated and lose their mechanical properties. Frankly no idea about plastics.
So, let's play the devil advocate here: how does this differ from the market for compiler tools on Linux? I am giving this example because once upon a time there was such a lawsuit by some lousy lamers with a pascal compiler against all major unix vendors. As a result they stopped shipping C without extra licenses. Somewhere in the late 80-es I believe.
There will be no real effort into curing their "deviations" for a long time to come because any attempt to do serious scientific research into it will get torpedoed by gay rights groups. Can't blame 'em. A cure for the urge to shag a 5 year old boy may end up being a cure for the urge of shagging any person of the same sex. Some of them do not like the idea of this cure being available because it is likely that it will be forced on at least some people. At least that is the official reason. The real reason is that there is no way to collect enough experimental data from sex offender cases while other "deviations"...
The only moderately successfull treatment has been stuffing them with chemicals to suppress any sexual urge. Scandinavian countries have been doing this for 10+ years now and have a nearly zero rate of pedofiliac associated crime (abduct, shag and kill cases) and nearly zero repeat offence rate. Other countries (France) are looking into it. Unfortunately neither UK, nor US have been even considering this. The obvious problem with this approach is that the moment the drugs stop the person is likely to reoffend.
So at least until we start really looking into the possibility of altering sexual behaviour via therapy the sex offenders are here to stay. If we do not lock them and do not stuff them with chemicals we might as well tag them. Not that it will help. It was considered in Sweden and France as an alternative to stuffing them with sexual depressants and it was found that when under the urge they do not give a flying fuck about any control measures. They will just do it, so the right way of dealing with it is to ensure they never get the urge in first place.
Out of these:
Reuters does not put all of its stuff on the web. It is a mixed subscription + free model.
Sun is not a newspaper. It is a form of porn disguised as news. It does not serve the same audience.
The Mirror is not a newspaper. It is a form of prudish porn disguised as news. It also does not server audience that plans to read news.
Times is planning to go to a subscription model. In fact, I think they are trialing a system for charging non-UK subscribers.
This in fact leaves only the Guardian and the Scotsman who are doing guess what? Campaigning for the extermination of news.bbc.co.uk as unfair competition.
There is a long and ongoing conflict between the BBC and the other online media in the UK. BBC has been able to nearly exterminate all UK online news except the specialized ones. Even those have nowdays moved to a subscription model (FT, DT and the like). It is free, it is high quality and it is updated round the clock. This is seriously annoying most newspaper owners and editors who see their circulation dropping and online presence being pushed into the oblivion. Allowing them to feed on an RSS feed is a great marketing ploy. They instead of protesting endlessly will now become largely dependant on BBC news online existing. Once this happens Blair (or any other creature which replaces him) will no longer have the option to exterminate news.bbc.co.uk for "efficiency of public spending" reasons.
Well... Most S.M.A.R.T. temperature sensors are on the PCB and they are measuring PCB temperature instead of the internal drive temperature.
Hence, a fan under the disk makes a lot of difference while making very little to make your data safer.
A 3x 40mm fan battery in front of a drive or a pressed enclosure that cools the actual package holding the platters makes a lot of difference there while not chaning the S.M.A.R.T. reading by more then a degree or so.
It is up to you - what do you want. Show (a good reading) or substance (good temperature of your drive platters and heads).
Err... Making nice with Mandelson? I may be wrong, but I do not recall Allen being the specific orientation essential for servicing Mandelson's orifices.
Not entirely correct.
There is also the bind8+ algorithm which is also followed by most other well behaved implementations. Besides the normal TTL criteria these will expire a record after every access with some non-zero probability. As a result records leave the cache before the TTL is reached.
This is done as a safeguard against typos in TTL and operational mistakes. Not everyone likes it. For example there was a long rant by DJB about why it sucks once upon a time. But it is good and nearly all popular implementations do it nowdays.
Complete and utter bollocks.
This has happened before. See for the discussions about LMcV and lmbench in the 90-es. In fact, the moment I saw that Linus has AGAIN selected a Larry McVoy tool my first thoughts were "Oh no, not another lmbench". I bet I was not the only one.
Considering that he is also known to be litigation happy I am not going to qualify his behaviour that time and this time. Just read the LKM on both occasions as well as some of his musings. They are selfexplanatory.
It definitely is. Since when do molluscs have tits and need a bikini top?
It was a mistake to agree to work with a company run by Larry McVoy in first place. Some of us remember previous McVoyisms from time past. Namely lmbench.
For a long time it was the best stress testing tool for Linux (and Unix for that matter). In fact if you look back around 1995 it was the de-facto standard for both performance and stress testing. That continued until Larry got pissed off at someone (forgot exactly what, look in lkm for the mid-nineties). So the tool went unmaintained and into oblivion.
There is no point to try working with people like this on any large project. They take stuff which is "just business" personally. They also get pissed off personally when someone finds a weakness in their way to do business and exploits it legally. So on so fourth.
My first reaction when BM surfaced 3 years ago was "Oh no... Not another lmbench". And I see that I was right...
That is correct. Another set of affected software which is not listed but affected is all opensource software using the old TAP/TUN driver. CIPE. OpenVPN 1.6, etc. All of these break. Interestingly enough some of them do not break immediately, but break sooner or later anyway.
1. Noisy - there is an environmental limit in the EU. It cannot be more than that. My car is fairly close to it and is tolerable on up to 5h trips so how to put it - do not really care.
2. Slow + short range - commuter car. Speed and range are not part of the spec. If it can move me the 5-10 miles to the office and back every day who cares.
Basically you are in the usual mistake of believing the top gear reviews. They are quite often valid, but suffer from the following serious deficiencies:
1. Car should fit Jeremy Clarkson in an American driving position. Despite being a "car guru" he still has not learned that you have to drive many cars in a french driving position (up and close to the wheel) to get the most of them. He drives splattered so he ends up not fitting in things like the Modus, Sirion, Copen, etc.
2. Car should require wrestling with it by Jeremy Clarkson to be driven. If it has a decent light electric power assisted steering it is blamed as lightweight, toylike and a few other adjectives.
3. So on so fourth.
Do not understand me wrong. They are still better then consumer reviews, but I would not trust them a lot. I would read carkeys or something a bit more balanced instead (where people who test know how to drive).
Show me a station offering free air at 300 bar. Wanna see one...
The basic problem with this car is that it will require extra infrastructure. Not terribly expensive, but quite noisy. Compressing air to 300 bar is not a very quiet affair.
Nope.
You are wrong.
They are getting better, but getting more narrow and more slanted towards a specific profession. For example the referred RFC is really funny only if you know some network engineering. If you do, you will end up in cramps on the floor from hysterical laughter. It is even better then RFC on IP over Avian Carriers with QoS and draft on Electicity over MPLS.
After all it is an IETF document - it is supposed to be funny to the IETF audience.
So what data does it have to write to disk if the bloody calendar is bloody empty? Similarly what data does it have to write if it is supposed to use a network calendar? So on so fourth. It is a bug (and filed as such).
I buy from these guys: http://www.a1-battery.co.uk They are a front-end for several Chinese and Korean manufacturers. If you are not in the UK get the part number from them and search in your country. For example their Sony equivalents have 25%+ charge for nearly all models. I have not tried IBM and others but I would expect them to be the same. The only problem I have found is that they sometimes switch from trickle charge to full charge mode (actually I see it with all new batteries, my phone does the same). This results in annoying notifications "your battery is now full" once in a while.
Ahem. Seconded.
:-)
I get 7+ hours out of an ancient Sony P3 1GHz bog standard desktop replacement class Vaio (3+ years old) under linux after throwing out the horrid puke sony ships for battery and replacing the battery and the CD with 2 x Chinese OEM 4800 mAh and using cpufreqd to keep the CPU frequency as low as possible when idling.
All I want is that the idiots at KDE HQ stop calling sync after each disk write operation to calendar and settings. I was almost ready to rebuild the entire thing with sync redefined as NOOP at one point. That will give me 1 more hour
But definitely, do not see any case whatsoever for using an external battery. If you need that charge level simply throw out your CD and line up 1 or 2 cheap and chearfull Chinese or Korean OEM batteries. They quite often have 20-30% higher charge then the branded stuff and cost twice less.
The most interesting screenshot as far as potential users are concerned is missing. I do not see the screenshot of the mail client. While I am not a particular fan of IE, it works to some extent so the users of Win CE have a working browser.
What they do not have is a working mail client.
The built in portable outlook only fetches recent messages and only from Inbox. You cannot access IMAP folders other then INBOX and have no means of accessing your old mail. To add to that, Pocket PC Outlook on the Smartphone Edition does not have client side SSL certificate support for IMAP or POP3 so you cannot provide secure remote access. This is essential, because it does not support any (even lame) encrypted VPNs and there are no third party clients for it.
There are plenty of people out there who will _pay_ for a working thunderbird on Windoze Smartphones. And very few who actually care about a working third party browser to replace IE.
Both are nice cards, but I would not recommend them to anyone who does not have extensive PC hardware knowledge. They are fussy, carpicious and very hard to troubleshoot when they go wrong.
One more point.
Most SpaceShip one type efforts employ a subsonic plane "booster" because of the separation problem. There is no airframe design out there capable of performing all of the following: horizontal takeoff, lift, flight and successful separation at supersonic speeds while in the atmosphere. You either separate at subsonic speeds like the Pegasus and similar Russian efforts or you do the full monty and get out of the lower layers of the atmosphere and separate there the way all missile launchers do (first stage is usually to 20km at least).
If someone designs a separable airframe the technology to achieve Mach 3+ and 20km+ altitude is already there. Mig 31 in the reconnaissance version can do it, BlackBird can do it and these are 20 year old designs. While there is 22 Mach to go, if you look at the fuel expenditure doing 22 Mach from 20+ km altitude is likely to be about 10% of the fuel compared to doing 25 Mach from 0.
And anyway, methinks that spaceplane booster stages are the wrong idea. I think it will be cheaper to build a rail launcher somewhere at 4-5 km altitude that fires a missile payload at 10-20 degrees up. Accelerating a few hundred tons to 600km+ for a very short trip is no big deal by modern technical standards. There are places in the Andes, Rockies and Pamir that are reasonably flat, non-inhabited and have suitable roads and infrastructure close by at altitudes as high as 4km. It is only a matter of time until someone decides to do this.
It is not a personal bug. It is the showstopper bug as far as any network engineer or IT person is concerned.
There is no way to do a useable presentation with a network diagram in it as far as OO is concerned. Same for any more complex workflow, database, UML, you name it. They are all vector graphics which makes OO unusable in the industry most likely to be willing to look into adopting it (we all hate microshit).
OO uses for preview only raster out of EPS. Some EPS have an embedded low res raster for previews, some don't depends on the converter. Vector and PS proper are not visualised. So while EPS sometimes it works for printing copies you:
1. Lose WYSIYG. You do not see your diagram in relation to the text. Which is about the level of "userfriendliness" I get from LyX or even TeX/LaTeX.
2. Cannot use it for electronic viewing. Most importantly cannot do impress/powerpoint presentations. This is how I ran into this. I tried to do a network diagram which came out fairly reasonable in one of the recent Dia versions. After that I found out that there is no way in hell I can get that diagram into any powerpoint compatible presentation editor.
Well, you simply have not run into the other big OO gremlin yet.
Not a single vector graphics import format works properly.
The ones that barely work (Autocad for example) lose colors and most of the formatting. So if you want to draw a half decent diagram using DIA and import it into an OO presentation you might as well forget it. Your only chance is to export it as a raster image and import it in OO. The result is horrible by all means. Horrible size, horrible visually, horrible in a print form and horrible to edit.
And OO 2.0 does not fix a single one of this issues. Instead of that we get visual candy - KDE widget support. Excuse me, but can we actually get the basic functionality fixed first before we get into Clippy land.