Insider preview of the 2013 Mac Pro design principles.

[flat five]

I'm not crazy enough to claim it's a "feature"

fwiw, i get how you guys are interpreting my use of the word 'feature'.. it's standard internet jokes which i tried to use in reverse and i understand how it hasn't gone over so well or worse - not be helpful to the point i was trying to make..

the point was that if the designers (again, not us) wanted bulk storage in the computer, it would be in the computer.

 

[ybz90]

the point was that if the designers (again, not us) wanted bulk storage in the computer, it would be in the computer.[/COLOR]

You're still not getting it. That is a logical fallacy because unless you're one of the nMP's designers, you can't know this to be true. Beating this example to death, there is no physical space left in the machine for bulk storage. To put it in there, the form factor that defines the nMP right now could not exist. They had to choose between the two and thus, the compromise was made to eliminate knowing that they had alternatives such as PCI-e SSDs and Thunderbolt that would give as good/better speeds compared to SATA. You keep talking as if you have authoritative knowledge on Apple's thought process, yet this as attained not with actual knowledge on the matter but by making assumptions on the final product.

I'm done with this thread too. Good day, folks.

 

[Rich.Cohen]

because that's not a compromise-- it's a feature.

"If you can't fix it, feature it."

Your example of no internal hard disk would be a constraint if the requirement stated that as an unalterable order. If the requirement was to make it really small and light then the elimination of addition drive space was a compromise. In any case all the "features" of a product are the result of either design compromises or management constraints. Without a copy of the requirements statement you can't tell them apart.

anyway- using 'compromise' as if it's a derogatory word then rebutting the phrase 'no compromise engineering' with it as if they're antonyms will just lead to miscommunication.

As an engineer I feel the compromise is something to be proud of. "You made some intelligent and imaginative compromises in satisfying my requirement."

My objection is that the phrase "no compromise engineering" will mislead non-engineers about a critical issue of engineer. I shudder to think of Dilbert's manager asking me as a consultant to deliver a "no compromise" solution. Actually I know my response - "Goodby. If you don't want compromise I don't want to do that for your and there is no compromise for me on this issue."

 

[ybz90]

As an engineer I feel the compromise is something to be proud of. "You made some intelligent and imaginative compromises in satisfying my requirement."

My objection is that the phrase "no compromise engineering" will mislead non-engineers about a critical issue of engineer. I shudder to think of Dilbert's manager asking me as a consultant to deliver a "no compromise" solution. Actually I know my response - "Goodby. If you don't want compromise I don't want to do that for your and there is no compromise for me on this issue."

If I could rep you somehow, sir, I would.

 

[Rich.Cohen]

The amount of complaining about the hard drive bays blows my mind. I'm not crazy enough to claim it's a "feature", because it's not

Actually it is a feature -- a characteristic of this design that distinguishes it from many other designs that compete to satisfy roughly the same set of user requirements.

Whether it is a good or bad feature depends on your own requirements.

Whether it was a compromise depends of the requirements given to the engineers. (It may have been a constraint.)

Whether it was a good compromise depends on whether you can come up with an alternative design that makes a different choice in this matter and delivers a better overall result in terms of the criteria the engineers were given by management. Such a judgement is likely to be very subjective.

 

[flat five]

You keep talking as if you have authoritative knowledge on Apple's thought process, yet this as attained not with actual knowledge on the matter but by making assumptions on the final product.

there is no physical space left in the machine for bulk storage. To put it in there, the form factor that defines the nMP right now could not exist. They had to choose between the two

so my authoritative knowledge is worthless but your authoritative knowledge is acceptable?
i think im getting the hang of this.

 

[Rich.Cohen]

1. There is no such thing as zero compromise engineering.
2. Compromises are neither good nor bad; I never said they were bad, which seems to be the crux of your (mis)interpretations. They are inherently both -- you give up something to get another.
3. The nMP is full of compromises. It is defined by compromising certain features to conform to the specific set of limitations resulting from the design and thermal form factor.

I agree.

And not having bulk storage is not feature.

I disagree. I think, for better or worse it is a feature. See my other posts.

Consumer preferences are irrelevant.

Far from it. (I hope.) But its the job of management and marketing to refine their understanding of future consumer preferences into the requirements given to the engineers.

There is a related issue I need to mention. A really good engineer has a clear view of what the final product will be like and how it will be used and how it will affect its users. In pursuit of that vision, the engineer may disregard some requirements. Sometimes this results in a truly outstanding product. Often a poor engineer uses this as an excuse to deliver an unsatisfactory product. During the design process it can be hard to tell which outcome is more likely.

But there is no doubt it's not compromise free, and there are certain users here proclaiming it to be the pinnacle of engineering.

This bothers me as well and it is why I've been participating in this thread.

 

[flat five]

1. There is no such thing as zero compromise engineering.

there is though.. there totally is

it's when the designer is completely satisfied (and rightfully so) with the way they encountered and/or solved ALL(within reason) problems during design. **

it doesn't happen very often (percentage wise) and very few people ever get to put their skills to true test with the type of dollar backing these guys had and seemingly the amount of freedom they had in the process.


**(edit- oh.. and obviously everything works as expected.. this one is still more/less an unknown with the nmp)

 

[Rich.Cohen]

If I could rep you somehow, sir, I would.

Thank you. I'm semi retired now, so I can better afford to express my principles.

I have often delivered less than the best I was capable of. Usually it was because I was given less time to work than I needed. My client was generally happy with the result while I was frustrated. The other recurring limitation were clients who expected me to gather requirement by reading their minds. "I don't have time to talk about all those details. You an expert at this. Just get to work."

Just curious, has anyone here read any of the following books?
Secrets of Consulting
The Design of Everyday Things
Exploring Requirements - Quality Before Design

Each in its own way is wonderful and for me was mind expanding.

 

[Rich.Cohen]

it's when the designer is completely satisfied (and rightfully so) with the way they encountered and/or solved ALL(within reason) problems encountered during design. **

In school engineering students are often given trivial problems for which there is one expected solution. It may be necessary to test them this way, but I think that it is one of the flaws of formal engineering education.

Even with an unconstrained budget, there are always tradeoffs and every tradeoff requires compromise. Just to name a few:
Size
Weight
Power required
Performance (many different metrics)
Manufacturability
Reliability
Ruggedness
Durability
Maintainability
Economy of operation
Environmental Impact during manufacture
Environmental Impact after disposal

A really inexperienced engineer may may decisions about all of the above without realizing it and claim no compromise. An experienced engineer knows better and has the scars and regrets to show for it. To this day I remember a software design compromise I made and lived to regret 30 years ago.

 

[ciaran00]

Two problems that didn't exist, that Apple didn't need to solve:
It's not small enough
It doesn't look like a jet engine turbine

Problems they ended up creating:
No storage expandability, no RAID
No internal PCIE cards
No consumer gpu upgrades
Crappier thermals
No PSU upgrades
No enduser repairability or part replacement
Proprietary (read: prohibitively expensive) upgrades

Nobody may use all of these upgrade paths but for most people having AT LEAST one of these upgrade paths was a consideration in buying the Mac Pro.

Yeah this is going to go well for them.

 

[flat five]

A really inexperienced engineer may may decisions about all of the above without realizing it and claim no compromise.

right. i get that.. the parenthesis in the statement- (and rightfully so)
was meant to imply it's legit and not just a whim of an inexperienced engineer.

 

[Rich.Cohen]

right. i get that.. the parenthesis in the statement- (and rightfully so)
was meant to imply it's legit and not just a whim of an inexperienced engineer.

I just don't buy that as a definition of zero compromise engineering. It's a good definition of good engineering, though because as you noted the design must satisfy more than just the engineer.

My problem is that I just can't stomach the notion of zero compromise when associated with engineering. It sounds like something a salesman might say to a gullible customer and I shudder. I shudder to think of people wanting a zero compromise product. I shudder to think of people fooled into believing they have purchased a zero compromise product.

I want a product that was properly manufactured in accordance with a well thought out and tested design. I want that design to reflect a good set of compromises against a requirement that closely approximates my requirements. I want the people who created that requirement to have included elements that address important issues that I don't know enough to ask for. I want a lot. I often don't get it. Sometimes I get it and don't know that I have.

About 35 years ago I was using some similar leading edge products from both HP and Tectronix. They were not oscilloscopes. On a number of occasions I needed to do something not in the users manual so I would call a sales engineer. The response from Tectronix was often "That's interesting. Let me get back to you." This was a professional response and they always followed up. The response from HP was often "Let me send you a white paper one of our engineers wrote on that." I don't know if that was better engineering or not, but being ahead of my needs certainly earned my respect.

I hope Apple is ahead of my needs with the nMP.

 

[flat five]

I just don't buy that as a definition of zero compromise engineering. It's a good definition of good engineering, though because as you noted the design must satisfy more than just the engineer.

My problem is that I just can't stomach the notion of zero compromise when associated with engineering. It sounds like something a salesman might say to a gullible customer and I shudder. I shudder to think of people wanting a zero compromise product. I shudder to think of people fooled into believing they have purchased a zero compromise product.

i do think i'm pretty clear on what you're saying here -or- i don't think i'm misinterpreting what you're saying.. and i agree with you completely..

it's more of us back and forthing from a semantics(?) pov on the phrase 'zero compromise engineering' .. another thing i agree with you on is the phrase carries a sort of cheesiness and it's not a choice of words i'd personally like to use either.

i know the feeling but i don't know the right words for it.. 'well engineered' almost works except it's lacking.. things can be well engineered even if the engineer wasn't giving her best effort..

but anyway, if someone happens to know the proper wording, can you please go back and edit miju's post in order to take the word 'compromise' out of it?

 

[Rich.Cohen]

it's more of us back and forthing from a semantics(?) pov on the phrase 'zero compromise engineering'

This is true. I'm a bit obsessed by semantics. I spend a lot of time trying to understand business requirements and I've learned the hard way the words can have different meanings to different people. I try to understand those meanings, but I also try to encourage people from muddying up a precise and well defined term. It sets my teeth on edge because it diminishes my ability to communicate. Hacker is an example. I used to be a hacker about 40 years ago. I once worked for someone who called any project diagram with a time line a PERT chart. Most of them were Gantt charts.

but anyway, if someone happens to know the proper wording, can you please go back and edit miju's post in order to take the word 'compromise' out of it?

Well designed is one phrase I'd use. An elegant design is another. An uncluttered design is yet another. All of these require a lot of subjective judgement. Perhaps an enduring design is the highest praise and one seldom found in computers. By this I mean a design that has proven its value over an extended period of time by satisfying a large segment of its intended user community.

Some examples of enduring design are:

The zipper
The DC-3
The AK-47

One might argue to include the SR-71, but it had a very limited production run. On the other hand, I would argue that the X-15 merits inclusion. Again there is a lot of subjective judgement here.

In the computer area the IBM-360 and the DEC PDP-8 might qualify. The DEC VT-52 might also fit the bill. Would anyone care to argue for or against the Apple-2?

 

[MiJuConcept]

Mr Cohen, as you know, when you have a square cooling tower and 3 sides are very hot and one side is very cold, you end up with horrendous amounts of tension - probably enough to warp the shape of the computer. The triangle doesnt have that problem as all 3 sides perform at working temps.

That is why the computer doesn't have one side of a four sided core dedicated to hard drives (Albeit PCIe storage or mSATA). A thermal mismatch in the cooling core is not a pretty thing.

The central thermal core allows the design you see now and it cannot be resolved with the amount of material used or same external dimensions. Calling that a compromise is an aesthetic argument - which I don't think engineers can overcome with brutal logic.

Designers dont speak of compromise anyway ... they only speak of balance. Pleasing the 99.9% to irritate 0.1%.

If you hate the MacPro ... you would really hate the MacBook Air experiment... the one with no screen, no keyboard and no battery. Just a motherboard wrapped in titanium.

 

[goMac]

No storage expandability, no RAID

Internal RAID in the Mac Pro has always been pretty useless anyway. You've got 4 drives max, and if you include redundancy (which you should), that's not going to get you far.

if you're a pro, you probably already have external storage of some sort, whether it's network storage or locally attached storage.

If you aren't a pro but you were buying a Mac Pro for the drive bays... well... think of this as a nudge back to consumer gear that might save you some money.

PSU upgrades? That never happened on the current Mac Pro.

There are at least a few components that are user replaceable as well.

 

[Rich.Cohen]

Mr Cohen,

Designers dont speak of compromise anyway ... they only speak of balance.

If you hate the MacPro ...

I'm an engineer. I've been designing things for about 40 years. I talk about compromises a lot. I don't use the term balance when I mean compromise, but I can understand some people might prefer that term. Most other engineers I know frequently talk about design compromises.

Where did you get the notion that I hate the MacPro?

 

[MiJuConcept]

Where did you get the notion that I hate the MacPro?

If ....

 

[cgk.emu]

Two ideas created the 2013 Mac Pro. The first was a history dating back to the original PowerMac Cube which highlighted the idea of vertical compact design and high performance coupled with near-silent high airflow. The second was a link between the Lockheed SR-71 Blackbird, the PowerMac Cube and Macbook Pro as examples of titanium craftsmanship in the laptops and a tradition in extravagantly expensive engineering. A reference is made to the power of supersonic jet turbines as an enduring symbol of unrivaled speed and power. Specifically, SR-71 is one of the American symbols for zero-compromise engineering and the ultimate example of design in balance with performance. In response to this concept, the specifications were maximized, the engineering extravagantly expensive, and the black paint refers to stealth.

As with the SR-71, Apple made a conscious decision with this generation to allow for expensive design and expensive manufacture. One of the driving factors in the design was the outer shell which was deliberately made seamless to demonstrate both engineering excellence and design purity.

Working with the concept of jet turbine. When the outer shell is removed, the internal components are both easily serviced and strongly resemble a jet engine with the cowlings removed. The main cooling fan and thermal core takes design cues from jet turbine intake blades and combustion chamber. All of the thermal dissipation for CPU/GPU/ Power Supply are reverted to the "central thermal core" which maximized the air-metal contact area and thermal mass of the heat sinks and solved the problem of containing heat to one region of the system without leakage into the surrounding void. Externally, the cluster of connectors also refers to the central mounting point of a jet turbine to the wing of an aircraft.

Early appraisal stated it as "a challenge worthy of Apple ......a computer only Apple could consider making (in terms of engineering cost and manufacture cost)."

It has been stated by Apple that significant engineering challenges were posed by the thermal core in relation to the cooling performance, fan efficiency and noise levels. This caused a delay of over 12 months to launch. Several engineering innovations of this design include the interconnects between boards, fan blades, the triangular cooling system, CPU/GPU sandwiched between PCB and chassis, overall size, zero-legacy technology and deep-draw single-piece outer shell.

The first flight of the SR-71 was December 22, 1964. Almost 50 years later, the 2013 MacPro pays homage to unrivaled engineering with a December launch. The Pratt and Whitney designation of J58 might find its way into some markings internally. Like the plane, the new Mac Pro will be built in North America.

Keep an eye on the launch date to see if they managed to align it with the maiden flight !

Footnote: Apple felt that taking onboard a difficult engineering task like this would ultimately pay off by forcing the competition to engineer the same solutions from scratch. That would take time and cost significantly more than any computer ever built.

engineering a desktop workstation is maybe, and i'm being generous, 1/10000000th as hard as engineering the SR-71. I fail to see even a hint of a link between the two.

 

[flat five]

engineering a desktop workstation is maybe, and i'm being generous, 1/10000000th as hard as engineering the SR-71. I fail to see even a hint of a link between the two.

it's probably not though.. not in the big picture at least

meaning engineering of the sr71 was a culmination of the lockhead engineers plus all(most/much) other engineering via humans up to that point.

the nmp is 40+ years newer..

it was harder for humans to arrive at -or- more engineering went in to today's computers than high flying spy planes..

edit-

(ie- if you put a million animals that are a little bit smarter than a pig on a deserted rock.. are they going to figure out how to fly fast first or- quickly process data?)

what's neat about that little scenario is that the computers, which are lagging in the race, will eventually (or are currently) allow the aircraft/spacecraft to fly faster -- whose engineering limits have already topped out without the assistance of computers.

 

[MiJuConcept]

engineering a desktop workstation is maybe, and i'm being generous, 1/10000000th as hard as engineering the SR-71. I fail to see even a hint of a link between the two.

- central cooling core akin to main combustion chamber of jet / ramjet
- interior board layout
- shape of blades on fan and exhaust
- round body similar to jet engine
- black paint (first black painted MacPro)
- high precision machining
- central mounting points for cables and power
- Launched into the 50th anniversary year of SR-71 maiden flight
- Built in USA
- High end engineering for computer
- Very expensive build for computer
- Outperforms all computers in SFF class (ITX size)

An aside .... signed by Apple team inside cover

The person who came up with this concept lived directly under the flight path of the Aurora which is the prototype for the SR-72. Having seen the Mach6 performance of the Aurora, it drove a fascination with the above concept. I dont think anyone is saying that the nMP is able to operate spy missions at 80,000 feet or uses JP7 fuel.

It is just a computer that breaks away from tradition in almost every design area. Destined for MOMA in New York....

 

[Rich.Cohen]

it's probably not though.. not in the big picture at least

meaning engineering of the sr71 was a culmination of the lockhead engineers plus all(most/much) other engineering via humans up to that point.

the nmp is 40+ years newer..

it was harder for humans to arrive at -or- more engineering went in to today's computers than high flying spy planes..

edit-

(ie- if you put a million animals that are a little bit smarter than a pig on a deserted rock.. are they going to figure out how to fly fast first or- quickly process data?)

what's neat about that little scenario is that the computers, which are lagging in the race, will eventually (or are currently) allow the aircraft/spacecraft to fly faster -- whose engineering limits have already topped out without the assistance of computers.

A good measure of the size of an engineering challenge is the number and difficulty of the unknowns that must be resolved by the design team. By that measure I'm willing to argue that the SR-71 presented a considerably greater challenge to its design team than the nMP did to its.

 

[flat five]

A good measure of the size of an engineering challenge is the number and difficulty of the unknowns that must be resolved by the design team. By that measure I'm willing to argue that the SR-71 presented a considerably greater challenge to its design team than the nMP did to its.

hmm.. i get what you're saying but i don't feel as if it's corresponding to what i was saying..

for clarity.. what i said has nothing to do with apple or its specific design team.

the only reason i wrote 'nmp' in that post was because it's the most modern desktop to date.


the edit question should make it clear which pov i'm coming from.. except i'm not poet. #

 

[MiJuConcept]

A good measure of the size of an engineering challenge is the number and difficulty of the unknowns that must be resolved by the design team. By that measure I'm willing to argue that the SR-71 presented a considerably greater challenge to its design team than the nMP did to its.

If the original design was based on Madonna's Bra it would have been reflective silver and conical ... as it only has to give insight into the design process and not equate with every process.