So, in June the marketing guy says to me "you should go to the Microsoft dev conference this year", and mind you this is one week after I've finally cut over to Linux as my primary OS.
Well. as I sit here reading Slashdot on my free Samsung win8 tablet, I have to say I'm impressed. This thing may not be an iPad but it sure is better than the android tab I brought with me!
For the most part, in the WLAN card business the drivers are provided by the chipset manufacturers.
The guys who build the cards (ODMs with names you've likely never hard of like Global Sun), at the behest of the OEMs (DLink, USRobotics, etc), take the drivers from the chipset manufacturer (Broadcom, TI, Atheros) and add pretty logos and graphics but change none of the functionality.
So, your argument that ODM A will build a card with chipset X, and ODM B will also go out and build a card with chipset X and steal ODM A's drivers doesn't fly. They both get the drivers from chipset X's manufacturer.
My GPRS latency tends to be much better than what you quote. From Chicago O'Hare connected over TMobile GPRS: Pinging google.com [216.239.39.99] with 32 bytes of data:
Reply from 216.239.39.99: bytes=32 time=1101ms TTL=242 Reply from 216.239.39.99: bytes=32 time=1061ms TTL=242 Reply from 216.239.39.99: bytes=32 time=1259ms TTL=242 Reply from 216.239.39.99: bytes=32 time=858ms TTL=242
I'm pleased as punch with cellular data service. That "how did I ever live without this?" feeling you first had after you started using WiFi, it's even more so being able to slap that PCMCIA card in the laptop and get a decent internet connection anywhere your cell phone works.
Re:My list of Delphi grudges
on
Delphi Turns 10
·
· Score: 1
4. Hard to separate GUI from business logic. The visual components expect being connected directly to DB components (TDataSource & TDataset). This is good for quickly snapping up a simple app, but soon it bites back as your app grows.
I started to run into this problem and quickly discovered stored procedures [google html translation of a good overview stuck in.doc format]. Stored procedures are a great way to encapsulate business logic so that the same data shows up across applications and reports. It also can help database query speed issues.
But that was many years ago when I was a data slave...
These "ideas" you speak of that come from Object Pascal... where are they? Can you name _any_ of them?
Properties
Someone else said it already, but it bears repeating.
*real* component based development, the uses clause, forms; there are so many great features to list about Delphi and Object Pascal. However, properties are the root of most of the coolness.
But as they say "Imitation is the most sincere form of flattery"
The OEM modules for machine to machine communications, directly comparable to this solution being pluged in the artical, are on their way.
EDGE is really starting to get a broad rollout, real 3G is much more expenseive for the carriers implement, EDGE is just an upgrade of existing GPRS basestations.
However, this movie is so obviously one-sided and defamatory that it should not be considered a fair and accurate documentary.
Why should a documentry be fair? Who judges fair?
I don't know why so many people have this altruistic view about documentries. No film can be objective. Think about "When Animals Attack" in comparison to a hypothetical documentry from Sea World showing Shamoo and her relationship with her trainer. Each shows killer wales, but in a different light.
Everyone has a bias, and it shows up in much more subtle ways that the filmaker's direct naration over images.
You hit the nail right on the head. These kinds of services are the Kinkos of electrical/mechanical fabrication.
A full service PCB fab house, like a full service printing shop, looks over your design to make sure it makes sense and nothing will get lost in the tranfer from design to implementation.
Customers are never happy when they spend $5k and what they get back there's a completely bonehead error- even if it is their own fault! From what I see, Kinkos and PCBExpress specialize in orders under $1k.
I was thinking of doing a similar sort of thing, but most of the machines in my house are laptops. I looked but couldn't find any cardbus GigE adapters. Only PCI.
This is a very timely artical. I just started taking a course online through Stanford's SCPD program. Last night I was dying with the slow pace of the lecturer, wondering if I could make it through a whole year like this.
They stream out video at 128 kbits for Windows media player. The NYT artical mentioned enounce 2x as a plugin for WMP that will let you speed things up. I tried it, but apparently SCPD's servers don't dole out the data fast enough and enounce catches up to the end of WMP's cache, pauses, refills the cache, then goes again at double time, only to hurry up and wait again.
I've fiddled with performance settings for 30 minutes to no avail and am about ready to give up.
Anyone used anything better to speed up WMP streams?
They'll never release source, too avant garde for an old school company.
However there is a good chance they'll have driver support for Linux in the September-December timeframe.
Whether the OEMs like DLink and SMC will buy the Linux Driver Developer Kit is another question.
Either way, Write to DLink strongly explaining to them that you'd like to see Linux drivers!
BTW, You might be able to find a better link on the contact page. I didn't see a better entry other than their Business Deveopment team.
Misled by the marketeers
on
802.11g Slows Down
·
· Score: 5, Informative
The rate on the box != the actual throughput you get.
Due to protocol overhead, backwards compatability overhead, physical environment, yada yada yada, you'll see varying throughput.
With current implementations of the draft solution mixed mode performance is *terrible*. 10 Mb/s mixed mode is an improvement. Right now your draft.11g solution is probably barely reaching 8 Mb/s in a mixed mode network and confusing the hell out of any.11b stations listening. You'd be silly not to upgrade the firmware as soon as they provide it.
The standards body hasn't throttled down.11g 's PHY level data rate; theoreticly 54Mb/s worth of info is still being spit out into the air. What they've done is added a little bit more overhead so that the.11g stations don't completely butt out the.11b stations.
Still, by the end of the summer you'll see throughput at 30 Mb/s in pure.11g, with 15 Mb/s mixed mode (without adversly affecting.11b stations). The leader of the pack should be Texas Instrument's chipset, hopefully to appear in DLink's newest 11g offering.
In a pure.11g network you won't get 54Mb/s but if you use TI's chipset you'll get throughput approaching 30 Mb/s.
>By the way, the actual speed of D-Link 614+ and 650+ is 11MBbps with 22Mbps throughput. It is more efficient than the "original" 802.11b
Actually it is a real 22Mb/s mode at the PHY level, but since TI (who makes the MAC/PHY chip that DLINK uses) got out manuvered in the 802.11 standards wars it had to offer it as a proprietary mode. That's what DLink markets as "2x" mode. If you're in an wireless network with another card that doesn't support "2x" mode it will fall back to 11Mb/s mode.
Actual throughput isn't near the advertised rate. The 11Mb/s or 22 Mb/s shown on the box is the theoretical throughput at the PHY level. Due to the overhead of the 802.11 standard and vagries of TCP/IP you'll see much less throughput.
In real world FTP throughput tests done at my office DLink cards averaged 6.2 Mb/s within 115 ft of the AP, while comparable cards ranged from 3.75 Mb/s (SMC) to 4.6 Mb/s (Cisco)
I developed apps with Delphi for 4 years and was completely sold that Borland had the best solution around.
At my new employment I had a project that needed to be cross platform. I was itching to use Kylix, however it wasn't due to be finished for another 8 months. Looking into the details I saw that they built the corss platform support on QT.
After checking out the QT tutorials I was immediatly hooked. QT is intuitive; I can't think of many other APIs I would grace with this description. In addition to being well thought out it has a superb implementation. I've been using it happily for 2 years.
The only thing I miss is the strong third party component community that Delphi/Kylix has. I'm a huge fan of "buy don't build". You can really put quality touches on your app by finding the right component someone else has already made.
You can already find out how seemless the Object Pascal/C++ integration will be, Borland C++ Builder
I've used BCB, and it's a quality product. It's not as clean as Delphi, but some PHB's I've seen get scared when you tell them you're going to write software in *shock* something other than C/C++.
When it comes to working with databases it's just hard to beat the usefullness and quality of the Borland database components. You just have to do so much less work when you use them.
This is your ammunition when users come and ask
why the wireless network is slower than the wired network with fewer users (preventing contention adds more overhead in wireless)
The right answer is: wireless networks are just plain slower than wired ones. Wired networks claim 100Mb/s access and wireless ones claim ~1/10 of that at 11Mb/s.
Actually CSMACA (as opposed to CSMACD the medium contention handling mechanism wired 802.3 networks use) really plays only a small part in the speed of.11 networks. It's the format of.11 physical layer packets and quality of current PHY layer chips which limits the throughput of most solutions to around 4 Mb/s.
New cards coming out from US Robitics using TI 802.11 silicon get consistent throughput close to 7Mb/s. Linksys also uses the TI ACX100 chipset, but doesn't have quite the marketing machine USR does.
If you need more speed you ought to check them out. Still not like a wired network but a hell of a lot better than 4 Mb/s.
I've used DriverX many times for PCI and paralell port access. They have a new product "DriverX USB" that I'm just starting to check out for an upcoming USB project. It sounds like just what you're looking for.
If just need to wiggle some bits I'd suggest using DriverX or its ilk to access the paralell port, but if you really want to play with USB it should be able to help you too.
Slashdot Mirror
So, in June the marketing guy says to me "you should go to the Microsoft dev conference this year", and mind you this is one week after I've finally cut over to Linux as my primary OS.
Well. as I sit here reading Slashdot on my free Samsung win8 tablet, I have to say I'm impressed. This thing may not be an iPad but it sure is better than the android tab I brought with me!
Ever transfer files from one computer to another with a crossover cable? It's really nice to do it wirelessly- hence ad hoc mode.
For the layman's overview of ad hoc mode check out this overview if you want the nitty gritty read the standard itself
It is a real mode. And would be quite usefull if chipset manufacturers bothered to implement it correctly and test interoperability.
For the most part, in the WLAN card business the drivers are provided by the chipset manufacturers.
The guys who build the cards (ODMs with names you've likely never hard of like Global Sun), at the behest of the OEMs (DLink, USRobotics, etc), take the drivers from the chipset manufacturer (Broadcom, TI, Atheros) and add pretty logos and graphics but change none of the functionality.
So, your argument that ODM A will build a card with chipset X, and ODM B will also go out and build a card with chipset X and steal ODM A's drivers doesn't fly. They both get the drivers from chipset X's manufacturer.
YMMV
My GPRS latency tends to be much better than what you quote. From Chicago O'Hare connected over TMobile GPRS:
Pinging google.com [216.239.39.99] with 32 bytes of data:
Reply from 216.239.39.99: bytes=32 time=1101ms TTL=242
Reply from 216.239.39.99: bytes=32 time=1061ms TTL=242
Reply from 216.239.39.99: bytes=32 time=1259ms TTL=242
Reply from 216.239.39.99: bytes=32 time=858ms TTL=242
I'm pleased as punch with cellular data service. That "how did I ever live without this?" feeling you first had after you started using WiFi, it's even more so being able to slap that PCMCIA card in the laptop and get a decent internet connection anywhere your cell phone works.
4. Hard to separate GUI from business logic. The visual components expect being connected directly to DB components (TDataSource & TDataset). This is good for quickly snapping up a simple app, but soon it bites back as your app grows.
.doc format]. Stored procedures are a great way to encapsulate business logic so that the same data shows up across applications and reports. It also can help database query speed issues.
I started to run into this problem and quickly discovered stored procedures [google html translation of a good overview stuck in
But that was many years ago when I was a data slave...
See the ungainly looking card in that picture? The wide part is just as wide as a PCMCIA card.
try an e-Mobile SIM card reader, http://www.kinforce.com/en/ArticleShow.asp?Articl
Since you can use this open source application it won't need to bother with the crappy software the sim reader comes with.
You can read my short comments regarding it in my blog
This looks like disinformation to me as well.
A poorly supported vote rigging story that is easily debunked leads takes away from the power of real evidence of vote fraud.
I'm going to repost this link to blackBoxVoting.org that someone else mentioned previously. Bev Harriss discusses in detail why this looks fishy.
>so the Madden series has been around every year for at least a decade.
I remember playing John Madden football on my Apple IIc with its sub 15 inch green monochrome monitor!
That must have been sometime between 1990-1994, before I got my brand spanking new 486-66 so I could play DOOM.
(and yes, I did have to trudge uphill both ways at 2400 baud)
These "ideas" you speak of that come from Object Pascal... where are they? Can you name _any_ of them?
Properties
Someone else said it already, but it bears repeating.
*real* component based development, the uses clause, forms; there are so many great features to list about Delphi and Object Pascal. However, properties are the root of most of the coolness.
But as they say "Imitation is the most sincere form of flattery"
You can 384Kbps, with an EDGE modem (like the Sony Ericsson PCCard).
The OEM modules for machine to machine communications, directly comparable to this solution being pluged in the artical, are on their way.
EDGE is really starting to get a broad rollout, real 3G is much more expenseive for the carriers implement, EDGE is just an upgrade of existing GPRS basestations.
Why should a documentry be fair? Who judges fair?
I don't know why so many people have this altruistic view about documentries. No film can be objective. Think about "When Animals Attack" in comparison to a hypothetical documentry from Sea World showing Shamoo and her relationship with her trainer. Each shows killer wales, but in a different light.
Everyone has a bias, and it shows up in much more subtle ways that the filmaker's direct naration over images.
You hit the nail right on the head. These kinds of services are the Kinkos of electrical/mechanical fabrication.
A full service PCB fab house, like a full service printing shop, looks over your design to make sure it makes sense and nothing will get lost in the tranfer from design to implementation.
Customers are never happy when they spend $5k and what they get back there's a completely bonehead error- even if it is their own fault! From what I see, Kinkos and PCBExpress specialize in orders under $1k.
I was thinking of doing a similar sort of thing, but most of the machines in my house are laptops. I looked but couldn't find any cardbus GigE adapters. Only PCI.
Anyone else find one?
This is a very timely artical. I just started taking a course online through Stanford's SCPD program. Last night I was dying with the slow pace of the lecturer, wondering if I could make it through a whole year like this.
They stream out video at 128 kbits for Windows media player. The NYT artical mentioned enounce 2x as a plugin for WMP that will let you speed things up. I tried it, but apparently SCPD's servers don't dole out the data fast enough and enounce catches up to the end of WMP's cache, pauses, refills the cache, then goes again at double time, only to hurry up and wait again.
I've fiddled with performance settings for 30 minutes to no avail and am about ready to give up.
Anyone used anything better to speed up WMP streams?
A good bit of OMAP documenation is already available on the web.
(An AC already posted this in reply to this comment, but I think it deserves more visability)
What do you think is lacking in this documentation? They give you CodeComposer, have a device driver writing guide, give you the instruction set, etc.
I haven't used these particular documents myself, but have been pretty happy with other TI documentation.
> Did anyone else read the subject and interpret ":=" as the PL/SQL assignment operator?
I read it as the Pascal assignment operator!
You are getting confused by the theoretical bit rate 54 Mb/s and the actual data throughput of 10-20 MB/s.
The poorly written artical you're thinking of is a muddled piece of work, as many of the the comments in the slashdot discussion explained.
They'll never release source, too avant garde for an old school company.
However there is a good chance they'll have driver support for Linux in the September-December timeframe.
Whether the OEMs like DLink and SMC will buy the Linux Driver Developer Kit is another question.
Either way, Write to DLink strongly explaining to them that you'd like to see Linux drivers!
BTW, You might be able to find a better link on the contact page. I didn't see a better entry other than their Business Deveopment team.
The rate on the box != the actual throughput you get.
Due to protocol overhead, backwards compatability overhead, physical environment, yada yada yada, you'll see varying throughput.
With current implementations of the draft solution mixed mode performance is *terrible*. 10 Mb/s mixed mode is an improvement. Right now your draft
The standards body hasn't throttled down
Still, by the end of the summer you'll see throughput at 30 Mb/s in pure
In a pure
The compatability
Actually it is a real 22Mb/s mode at the PHY level, but since TI
(who makes the MAC/PHY chip that DLINK uses) got out manuvered in the 802.11 standards wars it had to offer it as a proprietary mode. That's what DLink markets as "2x" mode. If you're in an wireless network with another card that doesn't support "2x" mode it will fall back to 11Mb/s mode.
Actual throughput isn't near the advertised rate. The 11Mb/s or 22 Mb/s shown on the box is the theoretical throughput at the PHY level. Due to the overhead of the 802.11 standard and vagries of TCP/IP you'll see much less throughput.
In real world FTP throughput tests done at my office DLink cards averaged 6.2 Mb/s within 115 ft of the AP, while comparable cards ranged from 3.75 Mb/s (SMC) to 4.6 Mb/s (Cisco)
At my new employment I had a project that needed to be cross platform. I was itching to use Kylix, however it wasn't due to be finished for another 8 months. Looking into the details I saw that they built the corss platform support on QT.
After checking out the QT tutorials I was immediatly hooked. QT is intuitive; I can't think of many other APIs I would grace with this description. In addition to being well thought out it has a superb implementation. I've been using it happily for 2 years.
The only thing I miss is the strong third party component community that Delphi/Kylix has. I'm a huge fan of "buy don't build". You can really put quality touches on your app by finding the right component someone else has already made.
-R
You can already find out how seemless the Object Pascal/C++ integration will be, Borland C++ Builder
I've used BCB, and it's a quality product. It's not as clean as Delphi, but some PHB's I've seen get scared when you tell them you're going to write software in *shock* something other than C/C++.
When it comes to working with databases it's just hard to beat the usefullness and quality of the Borland database components. You just have to do so much less work when you use them.
This is your ammunition when users come and ask
why the wireless network is slower than the wired network with fewer users (preventing contention adds more overhead in wireless)
The right answer is: wireless networks are just plain slower than wired ones. Wired networks claim 100Mb/s access and wireless ones claim ~1/10 of that at 11Mb/s.
Actually CSMACA (as opposed to CSMACD the medium contention handling mechanism wired 802.3 networks use) really plays only a small part in the speed of
New cards coming out from US Robitics using TI 802.11 silicon get consistent throughput close to 7Mb/s. Linksys also uses the TI ACX100 chipset, but doesn't have quite the marketing machine USR does.
If you need more speed you ought to check them out. Still not like a wired network but a hell of a lot better than 4 Mb/s.
I've used DriverX many times for PCI and paralell port access. They have a new product "DriverX USB" that I'm just starting to check out for an upcoming USB project. It sounds like just what you're looking for.
If just need to wiggle some bits I'd suggest using DriverX or its ilk to access the paralell port, but if you really want to play with USB it should be able to help you too.
Buy don't build!