[JuiciPatties]

I thought the whole purpose of being a software company was for them to build software to be used across all platforms (or at least all relevant platforms). As such, they would be bringing their software to both Android and iOS. Like how they have done this with BBM, I assume that eventually, the other apps they have produced will make it to iOS. However, as Conite has suggested, if you had to prioritize one, you would definitely go after the 88% of the market first.

[BGRS]

I thought the whole purpose of being a software company was for them to build software to be used across all platforms (or at least all relevant platforms). As such, they would be bringing their software to both Android and iOS. Like how they have done this with BBM, I assume that eventually, the other apps they have produced will make it to iOS. However, as Conite has suggested, if you had to prioritize one, you would definitely go after the 88% of the market first.

:-) Apparently not, they are very picky...And lets face it , they would never bring BBM to any platform if they had any choice. But agree, they have still lot to offer to the other platforms if not focused only to the android.

[conite]

I would not put all of their eggs in one basket...

You're assuming they don't have finite resources.

[BGRS]

You're assuming they don't have finite resources.

I agree, but isn't that the reason why they are always late, I mean behind dev incompetence? If you make something late you better don't even try doing it.

[Bbnivende]

I did a quick search on iMore to see if there was any chatter about the Hub and I did not find any.

I wonder if the time has come for the BlackBerry software company to change their name to just Berry or QNX or RIM at least on the consumer side .

[Prem WatsApp]

Interesting post by BlackBerry. Hope things are going well in the future... :-D

�   "BB10 dead?" - "Let's dance the Danse MacaBBRY! ... or is it..?" ;-D   �

[sorinv]

Is there any reason why you can't speak in specific terms instead of these vague "answers"? For example, can you say that Intel is losing money in fab and can't keep up with Moore's Law, so they're going to shutdown operations because it's not profitable anymore? "roadblock" can be anything.

Roadblock means that Moore's law is approaching its limit, which is predicted by the ITRS to be reached by 2021 when the physical gate length of the MOSFET will be 10nm.
Right now, in 14nm FinFET technology, the physical gate length is about 24nm.
http://spectrum.ieee.org/semiconduct...inking-in-2021

Besides the physical limits, there are economic and technical hurdles which partly explained why Intel has taken more years to move from the 14nm to 10nm node, than it took them to move from 22nm FinFETs to 14nm FinFETs.


Posted via CB10

[Chemy JMHT]

For it looks like they are about to move more aggressively to the Software area and leave a little alone the Hardware area, maybe what was told in another thread of joining forces with a software company or something like that.

[sorinv]

And these two fabs are? And they will dictate what "rules"?

The "vertical integration" you keep hinting at... who's going to do it and what will be their integration? Be explicit... or you're just hand waving and have no idea what you're talking about.

Please try to be polite if you are asking for information.

The only two advanced CMOS foundries left are TSMC and Global Foundries.
If you are a small startup and cannot prove that you have a lot of business and a lot of money to pay upfront (it takes more than 50M USD to develop a new hardware product in 14nm CMOS) no foundry will let you fab your product.

This is not that different from Qualcomm not willing to develop drivers for BlackBerry and presumably not willing to disclose enough details about the design of their chips to allow BlackBerry to develop the firmware drivers for Qualcomm's chips.

If software is so profitable and hardware not profitable, what makes you think that hardware companies like Qualcomm or Intel would not move to develop the software themselves since assembling a software team is cheaper (at least according to Chen) and requires a lower barrier of entry than developing a wireless transceiver or a 10nm finFET technology.

There is a lot of acquisition and vertical integration happening around this industry right now:
Intel acquiring Altera
Qualcomm acquiring NXP
Avago acquiring Broadcom
Siemens acquiring Mentor Graphics (an even more interesting integration between a system house and a CAD and design house).

All of these have happened in the last year and they will affect the IoT, 5G and autonomous navigation ecosystems.

Google, Apple, facebook, will have to partner, and some have already partnered, with foundries and may end up acquiring them if they are to continue to dominate with software. They are already into power stations: hardware.
You can develop software but somebody has to allow you to run it on their hardware.
You cannot run software virtually yet, just like humans cannot eat only virtually yet and still be alive.

Posted via CB10

[JohnKCG]

Yeah, how did that work out for them?

Even if vendors play Google's game though, it still doesn't stop them mucking about with things. Samsung Touchwiz anyone?

Well kindles are selling well afaik, so not bad I say, Touchwiz is like LG or Huawei UI's, a pretty crappy one actyaly


Posted via CB10

[cbvinh]

Have you used iphone's mail client? I have talked to at least 10 iphone users since I bought mine and they all use something other other then native mail app. Even Google mail client.

Follow-up question is whether enough iPhone users would spend $1/month for the Hub to justify development... For the Android version, BlackBerry is already developing for their own handsets, so the cost is already sunk in. For iOS, it would be a new cost.

[cbvinh]

Roadblock means that Moore's law is approaching its limit, which is predicted by the ITRS to be reached by 2021 when the physical gate length of the MOSFET will be 10nm.
Right now, in 14nm FinFET technology, the physical gate length is about 24nm.
Transistors Could Stop Shrinking in 2021 - IEEE Spectrum

Besides the physical limits, there are economic and technical hurdles which partly explained why Intel has taken more years to move from the 14nm to 10nm node, than it took them to move from 22nm FinFETs to 14nm FinFETs.

All this implies that there will be slower progress or that different methods/technologies would need to be used.

[cbvinh]

Please try to be polite if you are asking for information.

The only two advanced CMOS foundries left are TSMC and Global Foundries.
If you are a small startup and cannot prove that you have a lot of business and a lot of money to pay upfront (it takes more than 50M USD to develop a new hardware product in 14nm CMOS) no foundry will let you fab your product.

If you're a small startup, you wouldn't be contracting with fabs that produce in the volumes these companies require. Further, these companies probably do provide small scale services to small startups to foster future business.

This is not that different from Qualcomm not willing to develop drivers for BlackBerry and presumably not willing to disclose enough details about the design of their chips to allow BlackBerry to develop the firmware drivers for Qualcomm's chips.

Qualcomm is perfectly willing for the right price. BlackBerry isn't willing to pay the price, given the low volume of hardware sales. And of course Qualcomm isn't going to release their proprietary chip designs for BlackBerry to develop their own drivers.

If software is so profitable and hardware not profitable, what makes you think that hardware companies like Qualcomm or Intel would not move to develop the software themselves since assembling a software team is cheaper (at least according to Chen) and requires a lower barrier of entry than developing a wireless transceiver or a 10nm finFET technology.

There is a lot of acquisition and vertical integration happening around this industry right now:
Intel acquiring Altera
Qualcomm acquiring NXP
Avago acquiring Broadcom
Siemens acquiring Mentor Graphics (an even more interesting integration between a system house and a CAD and design house).

All of these have happened in the last year and they will affect the IoT, 5G and autonomous navigation ecosystems.

Google, Apple, facebook, will have to partner, and some have already partnered, with foundries and may end up acquiring them if they are to continue to dominate with software. They are already into power stations: hardware.
You can develop software but somebody has to allow you to run it on their hardware.

Software is profitable for BlackBerry. Hardware isn't profitable for BlackBerry. This is not a general statement that only software is profitable nor can this truth about BlackBerry be extrapolated to the electronics industry as a whole.

There have always been acquisitions and who knows what will become of them? Some may be bought for technology know-how, some to keep tech away from competitors, some to only acquire patents, etc. With large companies that are publicly held, they will forecast to their shareholders if they're going to vertical. It won't be secret. You won't wake up one day and all of a sudden, Apple announces they're going head-to-head with Intel in CPU manufacturing nor will you find Intel announcing they're head-to-head against anything Apple makes.

You cannot run software virtually yet, just like humans cannot eat only virtually yet and still be alive.

Is this supposed to be profound?

[DamianWarS]

The post is a bit of a tell revealing that since JC they always intended to phase out in-house devices. So despite people hanging on to this dream that if only a Z60 was released or a Passport 2 BlackBerry never intended to keep going. Since the words "software company" were uttered the plan the entire time was to get rid of hardware and they are on schedule. This is no surprise of course but it does make BB look a little more put together.

[sorinv]

If you're a small startup, you wouldn't be contracting with fabs that produce in the volumes these companies require. Further, these companies probably do provide small scale services to small startups to foster future business.



Qualcomm is perfectly willing for the right price. BlackBerry isn't willing to pay the price, given the low volume of hardware sales. And of course Qualcomm isn't going to release their proprietary chip designs for BlackBerry to develop their own drivers.



Software is profitable for BlackBerry. Hardware isn't profitable for BlackBerry. This is not a general statement that only software is profitable nor can this truth about BlackBerry be extrapolated to the electronics industry as a whole.

There have always been acquisitions and who knows what will become of them? Some may be bought for technology know-how, some to keep tech away from competitors, some to only acquire patents, etc. With large companies that are publicly held, they will forecast to their shareholders if they're going to vertical. It won't be secret. You won't wake up one day and all of a sudden, Apple announces they're going head-to-head with Intel in CPU manufacturing nor will you find Intel announcing they're head-to-head against anything Apple makes.



Is this supposed to be profound?

No. It's obvious, like the rest.

You are wrong about small start-ups.

As for the Qualcomm drivers for BB10, you agree with me and with the small start-up: if you can't afford to pay the big bucks you go to swim in the kiddie pool in knee-deep water, and you still drown. See BlackBerry.

The point is that Qualcomm would not disclose the details of the chip for BlackBerry to develop the firmware and drivers themselves.
Therefore Qualcomm is in a position to refuse or to charge a price that BlackBerry can't afford to pay because they can write software just like BlackBerry, but BlackBerry can't design the IC: advantage hardware (Qualcomm).


Posted via CB10

[sorinv]

All this implies that there will be slower progress or that different methods/technologies would need to be used.

Would need to be invented.

Posted via CB10

[JohnKCG]

No. It's obvious, like the rest.

You are wrong about small start-ups.

As for the Qualcomm drivers for BB10, you agree with me and with the small start-up: if you can't afford to pay the big bucks you go to swim in the kiddie pool in knee-deep water, and you still drown. See BlackBerry.

The point is that Qualcomm would not disclose the details of the chip for BlackBerry to develop the firmware and drivers themselves.
Therefore Qualcomm is in a position to refuse or to charge a price that BlackBerry can't afford to pay because they can write software just like BlackBerry, but BlackBerry can't design the IC: advantage hardware (Qualcomm).


Posted via CB10

This is the main reason of why BlackBerry 10 can't return

Posted via CB10

[stlabrat]

with all said, let's not forget apple hardware do make profit, with backing of A10 chip and integration of both software and hardware... The latest Macbook pro with possible upgrade path 3d xpoint, look like future profit in the plan (partner in place). Software defined radio still need back up hardware to support... another example... profit is in the integration and access to hardware if you do not have in house capability...that means you need someone in house at least know the field and have the knowledge to spec the hardware and capable of provide the 1st article test for the hardware/software integration before it become std (pre-market evaluation kits, the integrate provide performance data for field usage as benefit for the IC house)... or wait until the new technology has excess capacity - matured and have left over to sell after the primary player's order filled - do not expect to work at leading edge, just wait to the dog fight race to bottom price war at the time... profit will be slim or hard to come by...

[JohnKCG]

with all said, let's not forget apple hardware do make profit, with backing of A10 chip and integration of both software and hardware... The latest Macbook pro with possible upgrade path 3d xpoint, look like future profit in the plan (partner in place). Software defined radio still need back up hardware to support... another example... profit is in the integration and access to hardware if you do not have in house capability...that means you need someone in house at least know the field and have the knowledge to spec the hardware and capable of provide the 1st article test for the hardware/software integration before it become std (pre-market evaluation kits, the integrate provide performance data for field usage as benefit for the IC house)... or wait until the new technology has excess capacity - matured and have left over to sell after the primary player's order filled - do not expect to work at leading edge, just wait to the dog fight race to bottom price war at the time... profit will be slim or hard to come by...

Difference is, Apple has Cash to do that, BlackBerry, for the other side, is recovering from the hit that was BlackBerry 10

Posted via CB10

[sorinv]

And this is the latest news on Google partnering with Intel. You need hardware if you want to be the leader in software.
https://www.blog.google/topics/googl...33ef-163989973

Posted via CB10

[cbvinh]

No. It's obvious, like the rest.

You are wrong about small start-ups.

As for the Qualcomm drivers for BB10, you agree with me and with the small start-up: if you can't afford to pay the big bucks you go to swim in the kiddie pool in knee-deep water, and you still drown. See BlackBerry.

The point is that Qualcomm would not disclose the details of the chip for BlackBerry to develop the firmware and drivers themselves.
Therefore Qualcomm is in a position to refuse or to charge a price that BlackBerry can't afford to pay because they can write software just like BlackBerry, but BlackBerry can't design the IC: advantage hardware (Qualcomm).

If the small start-up were developing an *OS* that requires porting drivers and they wanted to run on Qualcomm chips, then they would need big bucks OR convince Qualcomm that their OS has the potential to get Qualcomm more business in the future. BlackBerry is already an established business with cash, so why would Qualcomm treat them like a start-up with an OS that has potential future business? The track record hasn't been good, so what's in it for Qualcomm?

Qualcomm doesn't disclose details because it's their trade secrets. BlackBerry doesn't invest in the drivers because BlackBerry doesn't think they'll recover their investment, not because it's cost prohibitive.

[cbvinh]

And this is the latest news on Google partnering with Intel. You need hardware if you want to be the leader in software.
https://www.blog.google/topics/googl...33ef-163989973

Did you even bother to read the article, much less the headline:

"Google and Intel announce strategic alliance to accelerate cloud adoption in the enterprise"

It's just Google and Intel collaborating on cloud services. Who knows where that's going to go? Google and Intel collaborated with Android on Intel and Intel eventually dropped the Atom processor that was running Android. I have a Dell Venue 8 7000 that will no longer be getting Android updates because this collaboration failed.

[sorinv]

Did you even bother to read the article, much less the headline:

"Google and Intel announce strategic alliance to accelerate cloud adoption in the enterprise"

It's just Google and Intel collaborating on cloud services. Who knows where that's going to go? Google and Intel collaborated with Android on Intel and Intel eventually dropped the Atom processor that was running Android. I have a Dell Venue 8 7000 that will no longer be getting Android updates because this collaboration failed.

I guess you haven't read further down about the search hardware and new tensorflow processors.
It's Google designing or co-designing with Intel specific hardware coding and libraries to speed up AI software. They can't do it unless they have intimate knowledge of the internals of the processors. They may actually even co-design hardware.

"
Kubernetes - Optimizing Kubernetes for Intel architecture to support a broad range of workloads. Intel is a major contributor to the Kubernetes ecosystem and enables enterprises to run OpenStack as a managed Kubernetes workload. Intel and Google engineers have already delivered code optimizations that have improved feature transparency on Intel architecture. This work is also expected to improve workload capabilities, including virtual network performance and prioritization of shared resources.

Machine Learning - Accelerating TensorFlow performance on Intel processors, for example by allowing TensorFlow to leverage all CPU cores and vector lanes for improved parallelism, integrating high-performance libraries such as the Math Kernel library (MKL) into TensorFlow, and optimizing memory allocation and data layer operations across a range of topologies."

By the way, Intel has the only advanced US-owned CMOS fab.
The rest are Asian owned: TSCM (Taiwan), Samsung (South Korea), Global Foundries (Middle East owned with fabs in US, Germany and Singapore).

The options are getting smaller if you want to secure your hardware.
You can never know for sure what the fab embeds in the mask set and in your IC if you do not fully own and control the fab, even if you designed the IC.

Posted via CB10

[cbvinh]

I guess you haven't read further down about the search hardware and new tensorflow processors.
It's Google designing or co-designing with Intel specific hardware coding and libraries to speed up AI software. They can't do it unless they have intimate knowledge of the internals of the processors. They may actually even co-design hardware.

"
Kubernetes - Optimizing Kubernetes for Intel architecture to support a broad range of workloads. Intel is a major contributor to the Kubernetes ecosystem and enables enterprises to run OpenStack as a managed Kubernetes workload. Intel and Google engineers have already delivered code optimizations that have improved feature transparency on Intel architecture. This work is also expected to improve workload capabilities, including virtual network performance and prioritization of shared resources.

Machine Learning - Accelerating TensorFlow performance on Intel processors, for example by allowing TensorFlow to leverage all CPU cores and vector lanes for improved parallelism, integrating high-performance libraries such as the Math Kernel library (MKL) into TensorFlow, and optimizing memory allocation and data layer operations across a range of topologies."

It only says that Google is optimizing their code on Intel hardware. No where does it say that Google is designing hardware with Intel, nor does it imply.

By the way, Intel has the only advanced US-owned CMOS fab.
The rest are Asian owned: TSCM (Taiwan), Samsung (South Korea), Global Foundries (Middle East owned with fabs in US, Germany and Singapore).

The options are getting smaller if you want to secure your hardware.
You can never know for sure what the fab embeds in the mask set and in your IC if you do not fully own and control the fab, even if you designed the IC.

Straying off topic, perhaps for a chance to bring up an article about IC masks being compromised... already known.

[sorinv]

It only says that Google is optimizing their code on Intel hardware. No where does it say that Google is designing hardware with Intel, nor does it imply.

Google has already designed special processors with Intel
https://en.wikipedia.org/wiki/Tensor_processing_unit

Also from wikipedia in the footnote for vision processors:

"Tensor processing unit, a chip used internally by Google for accelerating AI calculations."


Posted via CB10