Percent of Hardware Budget Devoted to COTS Hardware

 

Considerable attention and speculation has been given to the use of COTS hardware across different vertical markets, and whether this trend is expanding, remaining stable or declining. If the use of COTS is expanding, one would expect to see an economic benefit to its use – hence, a more important measure of COTS utilization would be reflected in the budgeted amount of COTS hardware as a percentage of total hardware cost.

In a recent 2011 survey of embedded developers (653 respondents), EMF asked respondents to report the percent of their total hardware budget that was devoted to COTS hardware.

Table I presents their responses according to vertical market. Whereas Aerospace/Avionics and Military had the highest response (these data reflect the percent of the COTS hardware budget compared with total hardware budget) Datacom and Electronic Instrumentation had a better than average response.

Percent of Hardware Budget Devoted to COTS Hardware
2011 EMF Survey of Embedded Developers
   
Industry Average 23%
Auto-Transportation 19%
Aerospace-Avionics 30%
Consumer Electronics 22%
Datacom-Networking 27%
Electronic Instrumentation 24%
Industrial Automation 20%
Medical 20%
Military 32%
Office Automation 14%
Telecom 23%

                   Table I

Table II presents developer responses according to chip architecture.

It is interesting to note that the budgeted percent of COTS hardware is consistent across all architectures, DSP and FPGA, but it is significantly larger for dual core and multi core developments. This might be due to the recent inclusion of multiple cores in embedded developments where the focus might be on software development within a mostly reusable hardware configuration. It will be interesting to see if this data is repeated in 2012.

 

 

Percent of Hardware Budget Devoted to COTS Hardware
2011 EMF Survey of Embedded Developers
   
Industry Average 23%
8-bit 21%
16-bit 22%
32-bit 24%
64-bit 23%
128-bit 26%
DSP 23%
FPGA 21%
Dual Core 28%
Multi-Core 35%

                             Table II

Information regarding the survey and data can be found at

www.embeddedforecast.com.

Survey data and the use of the EMF Embedded Dashboard used to compute these data can be seen at:

http://www.embeddedforecast.com/emfmip_videos.php

Give me some of That UK Venture Capital – They must be throwing it away

Never ask a barber if you need a haircut – Warren Buffet

The giggles keep coming – Atego has purchased another company (albeit a competent one – notwithstanding a questionable fit). This time it was Hi Rely, a service company that actually knows what its doing – although at a premium price.

I guess that in Jay Gambrel’s “Alice in Wonderland” dreams he believes that he’s going to match IBM’s 35,000 person service capabilities. Jay once told me that he’s a banker and that he doesn’t have to understand how the embedded world works.

I’ve known Vance Hildebrand for a long time – and I have previously recommended him and his organization. At least Hi Rely offers a useful service and has good people – good offerings and good people, however, do not necessarily create a positive and revenue producing combination. Time will tell.

Knowing Vance as a friend and colleague, I asked him why he sold to Atego – going in my mind from Upper East Side to downtown Newark. I asked if  a) he’s in trouble with his bookie, or, b) grandma needed implants to get her job back at Hooters. For the record, Vance loved my email to him and told me that he passed wine through his nose while laughing so hard. He offered to explain his willingness to be bought out by Atego after the May ESC conference. Vance has not responded to my emails since the wine-nose incident and we are on the lookout for a much too happy guy. Vance was paid in dollars (or Euros) and not in Atego stock – I told you that he is competent.

So Jay, please tell us who these more than generous VC folks are so that companies with good prospects and poor financial positions can gain access.

Jay, as a gesture of good will from this side of the puddle – and hoping that I can throw a bone to your financial backers, I’d like to introduce you to Gregg Miller of Oak Grove, Missouri.

So let me posit the following question: What differentiates the Atego acquisition strategy from Mr. Miller’s business? The answer is that neither seem to make sense – BUT Mr. Miller has made $20 million with his product he calls “Neuticles”.

Mr. Miller felt strongly that post-neutered dogs were too ashamed to show their face (or southern exposure) to other dogs, so he created fake testicles (Neuticles) that he sells for $100 a pair (not including surgery). To date he has sold over 250,000 Neuticles. Truth be told, I have a distant cousin who had Pekinese-sized Neuticles implanted in her Great Dane (we suspect nothing sinister) . The dog probably had emotional problem to begin with.

I guess that Neuticles is an easier sell than realtime Java

So Jay, don’t give up – if Mr. Miller can find gold with his “not of this planet” inspiration, so can you. Rumor has it that there is an embedded Unix company for sale.

Interpreting Embedded Survey Outcomes – Different Surveys Report Different Results

Surveys of embedded developers to establish market parameters regarding RTOS use must be careful to avoid unintentional bias. 

I have recently been contacted by senior executives from RTOS companies mentioning UBM’s survey of embedded developers that placed FreeRTOS as the most used RTOS for embedded use in 2011 (14% of respondents). FreeRTOS placed 3rd in the 2010 UBM survey. Year-over-year EMF data is at substantial odds with these findings.

EMF takes no position on how UBM conducts their survey or on their results. EMF data for the past two years have shown that FreeRTOS usage has garnered less than 1% of total developer responses. This encompasses over 1200 responding developers. This is a significantly lower response than the 14% reported by UBM. EMF has no idea of how this discrepancy came about. We stand by our data as our surveys and responses are restricted and carefully monitored. We make no judgment regarding UBM’s methods or results. We are responding to requests to report our findings.

In EMF’s 2011 Annual Survey of Embedded Developers, eleven thousand embedded developers were statistically selected and sent invitations to participate in the 2011 survey. Six hundred and fifty three developers responded to our invitation. PIN numbers were assigned to each request so that we could insure that only those invited participated in the survey – and that they could respond only once.

In 2011 developers reported using an in-house RTOS (20.1%), Android (19.3%), XPE (16.5%) and CE (15.9%). FreeRTOS was used by 0.9% of respondents. From our perspective, the suggestion that FreeRTOS use would exceed that of in-house, Android, XPE, CE, or VxWorks use is beyond any reasonable reality check.

How to Tell the Difference between Market Intelligence and Market Stupidity

Ignoring important business information in order to save money is like saving up sex for your old age. Warren Buffet

An archeologist was searching along the Amazon River when he stumbled onto a tribe of warriors. They were both shocked and surprised to see each other. The archeologist cried out “Lord, please save me for I am totally screwed”. A black cloud appeared and a loud voice cried out “you are not screwed – pick up the rock in front of you and kill the chief”. He picked up the rock and threw it hitting the chief in the head killing him instantly. The archeologist looked up to see 40 tribesmen coming at him with their spears aimed at him. He looked to the cloud and a loud voice said …

“OK – NOW you’re screwed”

Sometimes market advice seems to work that way.

A decade ago the merchant computer board industry was abuzz with CompactPCI (cPCI) forecast to replace the VME Bus. Market analysts were calling cPCI the two-billion dollar marketplace. Given that cPCI was controlled by three vendors (by definition a commodity marketplace), we called it the “zero-billion dollar marketplace”. Advocates told me that I was looking at the tip of the iceberg. I said that we were looking at the tip of the ice-cube!

The outcome was a no-brainer. commodity markets don’t grow as strong and as large as polyopoly markets (those that support a broad range of vendors). In addition, virtually every cPCI design was custom – not off the shelf. So who won? Not us – we were correct in our analysis, but no one bought our research. Other market research firms made a lot of money selling what vendors wanted to hear – but those vendors (Force Computer and Motorola Computer Group, among others) are long gone.

We never regretted our strategy – although we didn’t make any money in having the correct analysis. Ten years later, we are still here and doing fine.

 

The EMF Approach to Comprehensive and Reliable Market Intelligence

EMF’s Market Intelligence Program involves three steps

  • A comprehensive and statistically accurate survey statistically derived to insure randomness
  • An initial series of cross-tabs to provide an overview of the survey results
  • And an interactive Dashboard that enables you to explore the data set to  determine relationships in the data that can be essential to one’s efforts

No survey is adequate unless the data can be interrogated from a multitude of perspectives to establish relationships and correlations. EMF does this using a unique tool we call the Executive Dashboard.

 

A well constructed survey and the use of the Dashboard can provide the following:

 

Determine comparisons between your competitor’s products and yours

• Determine developer metrics: number of developers per project; number of lines of written code as well as total lines of code; cancellations; designs completed ahead of or behind schedule – and how many months behind schedule; and, comparisons between pre-design expectations and final design results – and be able to do this for any vertical market, any chip architecture used, etc.

• One can look at developers’ most pressing concerns, what design processes are used and what developers believe are best practice

• Look at product line deficiencies and needs

• Look at what developers are planning to use and do

• Degree of satisfaction of customers with products and tools

• Find market messages that resonate with potential customers, and keep products aligned with these benefits

We have put together a series of videos to illustrate how our surveys are constructed and how the dashboard is used. We encourage you the reader and embedded professional to think about how you evaluate information that is critical to your success.

You don’t want that black cloud telling you that “now you’re screwed”. The only thing worse would be if you had paid for that information.

2011 Embedded Developer Design Outcomes – Comparing EMF Blog Reader’s Responses with Those of the Embedded Industry

 

This is not an April Fool’s joke. The fresh snow on my Massachusetts lawn and driveway was our unexpected April Fool’s joke. If you responded to our request of our reader’s to take the same survey as did our embedded industry respondents, be proud and go ahed and ask your boss for  raise.

We recently posted a Blog comparing EMF Blog readers with their industry counterparts regarding how they stay knowledgeable about embedded products and services. There was, as might be expected, a substantial difference between the two groups. Our Blog readers were far more aggressive is gaining information than their counterparts.

In that Blog we posited the question whether such professional conduct would provide better and more cost efficient design outcomes.

For the past 15 years, EMF has conducted detailed and extensive surveys of embedded developers. Using the EMF Executive Dashboard (a data analysis tool unique to the embedded market intelligence community) we have been able to relate the decisions of embedded developers to their design outcomes. This also enables us to develop ROI and total cost of development calculations to establish the most cost effective of the many possible design undertakings. Our readers and downloaders of our many white papers are familiar with our results.

In our 2011 EMF Survey of Embedded Developers, we asked our readers to take the same survey as that of the statistically accurate industry wide survey. It has been an assumption/theory of mine that embedded professionals that take the time to research and read authentic market intelligence would produce better and more cost effective design outcomes.

We have examined this comparative data and we present it in Table I.

       
    Ind ave EMF Bloggers
  Devel time Months – Start to Ship 13.9 12.5
  % behind schedule 47.0% 38.1%
  Months behind 3.8 3.8
  % cancelled 11.2% 10.9%
  Months before cancellation 4.4 3.7
  SW Developers/project 14.7 9.7
  Total Developers/project 14.7 9.7
  Average Developer months/project 204.3 121.3
  Developer months lost to schedule 26.3 14.0
  Developer months lost to cancellation 7.2 3.9
  Total developer months/ project 237.8 139.2
  At $10,000/developer month    
  Average developer cost/project $2,043,300 $1,212,500
  Average cost to delay $262,542 $140,437
  Total developer cost/project $2,305,842 $1,352,937
  Advantage   70.4%

 

                                                    Table I

It is clear from Table I that the EMF Blog readers (that took the time to complete the survey – congratulate yourself if you did and show this to your boss) experienced substantially lower design costs that that of their industry counterparts. One might assume that professionals that take the time to keep current with embedded information would experience better results.

Judge for yourself.

Survey respondents were asked “How close was your final design outcome to your pre-design expectation?” The choices made available were: within 10%, 20%, 30%, 40%, and 50% and not within 50%. EMF believes that design outcomes within 30% represent a good design outcome – and within 20% represent an excellent design outcome. Table II presents the comparison between the groups.

         
    Ind Ave  EMF Bloggers  
         
  Performance 67.5% 67.8%  
  Systems Functionality 70.6% 71.4%  
         

                                              Table II

There is no difference between the groups. What we might deduce from Tables I and II is that although the design choices by our Blog readers were enlightened enough to produce a significant savings in design costs, their ultimate design outcomes were comparable. Of course, the pre-design expectation of the Blog readers might have been higher that that of the industry at large. This is pure speculation – the data does not support the assumption.

Nevertheless, if I was the CEO I’d go for the significant savings.

How Embedded Developers Research and Stay Knowledgeable About Embedded Products and Services

 

Comparing EMF Blog Reader’s Responses with Those of the Embedded Industry

 

There is an old Chinese curse that says “May you live in interesting times.” As the pace of technology change continues to accelerate (did anyone consider the ascendancy of Linux in the past or Android today?) the plight of many embedded developers and managers (not to mention cost controlling CFOs) – many facing limited windows of opportunity – makes choosing among the best alternatives very difficult.

 

For the past 15 years, EMF  has conducted detailed and extensive surveys of embedded developers. Using the EMF Executive Dashboard (a data analysis tool unique to the embedded market intelligence community) we have been able to relate the decisions of embedded developers to their design outcomes. This also enables us to develop ROI and total cost of development calculations to establish the most cost effective of the many possible design undertakings. Our readers and downloaders of our many white papers are familiar with our results.

 

In our 2011 EMF Survey of Embedded Developers, we asked our readers to take the same survey as that of the statistically accurate industry wide survey. It has been an assumption/theory of mine that embedded professionals that take the time to research and read authentic market intelligence would produce better and more cost effective design outcomes.

 

We are currently examining this comparative data and we will report on our analysis in a later blog or white paper. The first step, however, was to look at whether there was a difference in how our readers stay knowledgeable as compared with developers in the industry at large.

 

Comparative responses to the question “How do you typically first become aware of and stay knowledgeable about embedded products and services?” is presented in Table I.

  Read the rest of this entry »

2010 Retrospective of Embedded Developer Choices and Market Trends

 

When we look at embedded developer’s issues – what they are doing, what issues are the most disconcerting to them in their embedded developments, and what we can take from changes that we have seen over the past several years – these stand out as emerging trends in an expanding embedded marketplace.

 

In summary:

 

  • There has been a shift in the RTOS marketplace from the traditionally powerful RTOSes (e.g., Integrity, VxWorks and LynxOS) to the smaller, highly efficient RTOSes (e.g., ThreadX, Micrium, Nucleus and MontaVista Linux).

 

  • Simulation-Modeling (e.g., Simulink, Rhapsody) tools are being used effectively in more designs, driven by code reuse, reporting of application software under conditions of changing of underlying hardware, and by the financial advantages that accrue.

 

  • Communication middleware use is changing from RYO to commercially available middleware (e.g., RTI) as the difficulties (and cost of maintenance) of network expansion proved unmanageable.

 

  • The use of testing tools is expanding in embedded applications including requirements and change management, validation and verification tools and static and dynamic testing. Companies such as LDRA are successfully integrating many of these tools into a single offering that is interoperable with DOORS and Rhapsody. EMF believes that integration with other tools sets – rather than stand alone tools – is the future.

 

  • The use of dual core and multicore processors for embedded designs is growing rapidly – notwithstanding the lack of excellent tools (particularly for symmetric multiprocessing). This need will create a new and substantial addition to the development tools marketplace.

 

It will be fascinating to analyze the results from the 2011 EMF Survey of Embedded Developers to see which trends continue, what markets appear to be in decline and whether new and important market insights appear.

 

As always, we welcome your comments and insights (jerry@embeddedforecast.com)

Comparing Embedded Design Outcomes With and Without Model-Based Design

Happy New Year to our readers

Jerry & Dolores Krasner

 

Announcing our first webinar of 2011 with The MathWorks All are welcome and will receive a complimentary copy of our report.

 

The landing pages for our webinar on Jan. 12th  2011 are now “live” on the website and collecting registrations:

 

9:00am session: http://www.mathworks.com/company/events/webinars/wbnr53412.html?seq=1

 

2:00pm session: http://www.mathworks.com/company/events/webinars/wbnr53412.html?seq=2

Opportunities and Threats Confronting Mobile Device Manufacturers and Application Developers

 

EMF has recently published a detailed analysis of the Mobile Device Industry focusing on best practices and ROI analysis. This paper is available for free download from our website.

 cellphone

  Overview: It’s not easy being a mobile device manufacturer today

 

 Profound economic conditions spawn profound opportunities and challenges. As in any economic contest, there will be winners, losers and those that can’t tell the difference. Market uncertainties challenge vendors and OEMs alike to find new niches, competitive advantages and markets that will remain stable while enjoying growth potential.

 

For example, the FCC has recently adopted a regulation that could dramatically improve our wireless devices. The rule offers a brand new and much-improved slice of the radio space for unlicensed use. The new frequencies are known as “white spaces” and result from making available to the public the frequency spectrum previously used by analog TV transmissions. This spectrum will enable greatly enhanced WiFi capabilities which will provide for mile-long connectivity as well as increased and more reliable data handling capabilities. One can imagine the many new and powerful apps that will emerge – most of which will require reusing developed code.

 

EMF believes that this will both increase opportunities as well as risk for mobile device manufacturers. EMF strongly believes that the issues we present in this report are essential to those manufacturers that will dominate this emerging market opportunity. Merely maintaining a growing position in the current expanding marketplace will not be competitively sufficient.

 

As processor costs continue their downward spiral, and as fickle consumers of mobile technologies are increasingly ready to move to more interesting and useful displays and functionalities, consumer-based markets are becoming characteristically disposable. New platforms appear nearly monthly – like quantum particles appearing and disappearing into the void of consumer insanity – characterized by more functional and entertaining GUIs that enable the technologically disadvantaged to easily link to social sites, read email, text and perform acts of information magic that would have made Merlin give up his wand.

 

Points of Pain: issues that mobile device manufacturers must consider to remain competitive

  Read the rest of this entry »

Establishing a value for Market Intelligence? Can you avoid the Coming Embedded Tsunami?

We all can agree that good market information is essential for a company’s growth, competitive and strategic planning and sales support.

 Giving up

 

        

       But what is worse – bad information or no information whatsoever?

 

 

 

 

 

 

 

 

I recently read a story about an explorer looking for artifacts along the Amazon River when he happened upon some 50 or more tribesmen each carrying spears. Fearing for his life he cried out “Dear Lord I’m totally screwed”. A dark cloud appeared and a loud voice cried out, “you are NOT screwed – pick up the stone in front of you and kill the chief.” The man picked up and threw the stone hitting the chief in the head, killing him instantly.

The explorer looked up to see more than 50 tribesmen running at him aiming their spears at him. Looking up to the dark cloud, he heard the voice say “NOW you’re screwed!”

 

This is what Dolores and I ponder when we are publishing survey-based market intelligence and using it to address opportunities and threats that confront our subscribers. Giving advice is easy – giving “good” advice based on statistically accurate comprehensive surveys is essential.  

 

We carefully track what we and what others publish and how close each of us comes to correctly forecasting the marketplace. As Dolores is relatively new to our industry, she is finding it hard to understand why market research groups that have been grossly wrong made much more money than we did. Dolores was beside herself asking what happened to those companies that got it wrong. The answer is nothing – newer companies bought their research and the analysts that were wrong found new lines of work. We, on the other hand, have been here for 15 years and plan to be here a lot longer.

 

Before we get to our forecasts, let me provide some background. Perhaps some of you can help me answer some of Dolores’ questions – questions that I wish I could answer. Times have changed since I was on your side of the industry.

 

 A bit of background:

 

Dolores moved from education to the dreaded private sector when we became engaged some 5 years ago. Though having sufficient credits to complete a Ph.D. in performance analysis for validating student progress (skills that would bide her well at EMF), she was already at the highest pay level she could attain and she was more interested in seeing student improvement first hand, rather than theorizing about it and writing papers for her peers.

 

The professional environment in which she performed was union-based and raises were predicated on time on the job, degrees and post-graduate credits – actual performance was not a consideration and tenure was earned in as little as 3 years.

 

Her transition to our performance-based world was easy – she studied hard and attended many conferences and technical presentations over the years, yet her background in data analysis and performance outcomes supported her curiosity regarding how businesses grow or fail based on how they use available information for strategic planning, competitive analysis and sales support. Not only was she able to support my data-mining efforts with her experience, but she is able to ask embarrassing questions of me when she read (and schoolmarmed) my reports.

 

As embedded professionals, perhaps you can relate to the questions she has asked – and maybe provide us some feedback.

 

It usually begins with “I’m sure that this is a stupid question, but …” If I can’t explain it to her, then I probably didn’t think it through sufficiently.

 

Here are some of the many questions that she asks about companies in the embedded industry that I have a hard time answering:

 

1)     Why do companies with very smart people spend hundreds of thousands of dollars on sales development and much less on marketing?

2)     Why to companies that understand the value of marketing for sales and competitive positioning – and spend in some cases millions of dollars in their marketing efforts – spend very little on detailed, relevant and comprehensive market intelligence information?

3)     Why do these very smart people choose to ignore clear market signs (even when made aware of them) that can significantly impact current markets, future revenue streams and company survival?

4)     Why do so many European-based company CEOs micromanage their executives instead of hiring the best and freeing them up to success? Is this a better approach to corporate controls?

 

Forgive what might be construed as a commercial – what really drives her buggy is that she asks me (and not infrequently potential subscribers) “Tell me what I’m missing here; Companies can have access to detailed data that they can use for sales promotion, strategic planning, and competitive analysis for less than one-fourth what they pay for an administrative assistant. Moreover, they can have access to the information that the US military and many prime contractors have about them in order to better respond. Why isn’t someone fired?”

 

 

 What’s on the Embedded Horizon?  EMF’s look at markets that are in transition – creating opportunities as well as threats

 

Looking at Mobile Devices and Consumer Electronics

 

If you are in – or looking to get into – this market segment you will find plentiful opportunities and incredible challenges. New processors and new platforms are appearing regularly and the fickle user base will drop you without hesitation. Now that analog TV signals have gone the way of the Platypus, the FCC is assigning these “white channels” for public use. Herein we will see long range WiFi, and enhanced data handling capabilities that will provide new and dazzling apps that will drive newer markets. Time-to-market will be a crucial consideration, so the ability to reuse code and to apply such code to new interfaces and hardware will determine who will win and who will lose.

 

Do you provide mobile or consumer electronic products or applications? Do you know what processors, OSes and platforms provide better and timelier design outcomes? Does it matter to you? Is your resume up to date?

 

Looking at Medical Devices

 

If we separate medical devices into two cadres; one that supports devices that are attended to by medical staff (ultrasound, CAT scans, renal dialysis, etc.), the other that needs to operate independently of medical support staff (patient monitoring) we see that the same technology that keep military and commercial aircraft operating under secure and mission critical conditions is overkill for medical applications.

 

Did you know?

 

  • The medical device marketplace has been growing at a double digit rate – and new considerations should enhance opportunities for embedded vendors in 2011 – IF they take the time to correctly understand the selling points
  • Considerable attention has been given to the medical marketplace by vendors seeking a safe haven from the expected decline in mil/aero. EMF believes that many such vendors don’t understand the market segment, how to sell to it, or what the users need and will purchase
  • It would be funny if it wasn’t sad that certain leading RTOS vendors are pushing their certified high power, mission critical OSes towards an industry that develops products whose defining frequency requirements are less than 100 Hz
  • New CDRH/FDA initiatives are acting to create opportunities and threats.

 

Having brought many products through the 510k process and to market, and understanding the purchasing value system within the medical device marketplace, I am amused at the hundreds of thousands of dollars that vendors throw away on pursuing meaningless approaches to this marketplace. Dolores gets the “everything looks like a nail theory” and it was hard for her to keep a straight face when one colleague who is a vendor’s VP told us that it took them a year (and mucho dinero) to discover that what we gave them for free was correct.

 

The Coming Tsunami

The Aerospace and Defense Industries of the United States are poised to undergo one of the most significant changes since the end of the Cold War; perhaps the most significant since World War II. We believe that observers (embedded vendors included) who expect small changes are mistaken, thereby fostering a false and dangerous sense of security across much of the industry and government.

 

The impact to our economy goes far beyond our current financial problems and involves fundamental structural changes taking place in the industry and in the market. As in prior shifts of this nature, there will be winners and losers – however these shifts may be profound, creating more dramatic winners and losers than in the last cycle.

 

The impact of major primes shifting financial challenges to their vendors (and by association to embedded sub-vendors) will reach down into the value chain, in some cases devastating naïve second and third tier vendors.

 

To make matters more challenging for industry, as spending draws down, the effect on industry will be dramatic beyond the proportions of top line reductions.  This will be caused by a squeeze on the so called “Investment Accounts.”  These Investment Accounts are the funds used for product development and buying equipment.

 

Embedded vendors that offer “high end” OSes and tools will be hardest hit as reductions in DoD discretionary funding and the actions of prime contractors to absorb overhead within the limits of their financial structure.

 

We expect to see reductions in outsourcing to and purchasing from embedded markets defined as “mission critical” and that involve MILS security, DO-178B, and virtualization technologies as these anticipated contractions take place over the ensuing 2-3 year period. Many high-end vendors are already looking to alternative markets to shore up expected reductions – but most lack the data to determine which markets they can competitively serve.