Sunday 14 July 2013

Slower Moore's law wouldn't be that bad.

Many aspects of the world of computing are dominated by Moore's law -- the phenomenon that the density of integrated circuits tends to double every two years. In mainstream thought, this is often equated with progress -- a deterministic forward-march towards the universal better along a metaphorical one-dimensional path. In this essay, I'm creating a fictional alternative timeline to bring up some more dimensions. A more moderate pace in Moore's law wouldn't necessarily be that bad after all.

Question: What if Moore's law had been progressing at a half speed since 1980?

I won't try to explain the point of divergence. I just accept that, since 1980, certain technological milestones would have been rarer and fewer. As a result, certain quantities would have doubled only once every four years instead of every two years. The RAM capacities, transistor counts, hard disk sizes and clock frequencies would have just reached the 1990s level in the year 2000, and in the year 2013, we would be on the 1996 level in regards to these variables.

I'm excluding some hardware-related variables from my speculation. Growth in telecommunications bandwidths, including the spread of broadband, are more related to infrastructural development than Moore's law. I also consider the technological development in things like batteries, radio tranceivers and LCD screens to be unrelated to Moore's law, so their progress would have been more or less unaffected apart from things like framebuffer and DSP logic.

1. Most milestones of computing culture would not have been postponed.

When I mentioned "the 1996 level", many readers probably envisioned a world where we would be "stuck in the year 1996" in all computing-related aspects. Noisy desktop Pentiums running Windows 95s and Netscape Navigators, with users staring in awe at rainbow-colored, static, GIF-animation-plagued websites over landline dialup connections. This tells about mainstream views about computer culture: everything is so one-dimensionally techno-determinist that even progress in purely software- and culture-related aspects is difficult to envision without their supposed hardware prequisities.

My view is that progress in computing and some other high technology has always been primarily cultural. Things don't become market hits straight after they're invented, and they don't get invented straight after they're technologically possible. For example, there were touchscreen-based mobile computers as early as 1993 (Apple Newton), but it took until 2010 before the cultural aspects were right for their widespread adoption (iPad). In the Slow-Moore world, therefore, a lot of people would have tablets just like in our world, despite the fact that they wouldn't probably have too many colors.

The mainstream adoption of the Internet would have taken place in the mid-1990s just like in the real world. 1987-equivalent hardware would have been completely sufficient for the boom to take place. Public online services such as Videotex and BBSes had been available since the late 1970s, and Minitel had already gathered millions of users in France in the 1980s, so even a dumb text terminal would have sufficed on the client side. The power of the Internet compared to its competitors was its global, free and decentralized nature, so it would have taken off among common people even without graphical web browsers.

Assuming that the Internet had become popular with character-based interfaces rather than multimedia-enhanced hypertext documents, its technical timeline would have become somewhat different. Terminal emulators would have eventually accumulated features in the same way as Netscape-like browsers did in the real world. RIPscrip is a real-world example of what could have become dominant: graphics images, GUI components and even sound and video on top of a dumb terminal connection. "Dynamic content" wouldn't require horrible kludges such as "AJAX" or "dynamic HTML", as the dumb terminal approach would have been interactive and dynamic enough to begin with. The gap between graphical and text-based applications would be narrower, as well as the gap between "pre-web" and "modern" online culture.

The development of social media was purely culture-driven: Facebook would have been technically possible already in the 1980s -- feeds based on friend lists don't require more per-user computation than, say, IRC channels. What was needed was cultural development: several "generations" of online services were required before all the relevant ideas came up. In general, most online services I can think of could have taken place in some form or another, about the same time as they appeared in the real world.

The obvious exceptions would be those services that require a prohibitive amount of server-side storage. An equivalent of Google Street View would perhaps just show rough shapes of the buildings instead of actual photographs. YouTube would focus on low-bitrate animations (something like Flash) rather than on full videos, as the default storage space available per user would be quite limited. Client-side video/audio playback wouldn't necessarily be an issue, since MPEG decompression hardware was already available in some consumer devices in the early 1990s (Amiga CD32) and would have therefore been feasible in the Slow-Moore year 2004. Users would just be more sensitive about disk space and would therefore avoid video formats for content that doesn't require actual video.

All the familiar video games would be there, as the resource-hogging aspects of games can generally be scaled down without losing the game itself. It could even be argued that there would be far more "AAA" titles available, assuming that the average budget per game would be lower due to lower fidelity requirements.

Domestic broadband connections would be there, but they would be more often implemented via per-apartment ethernet sockets than via per-apartment broadband modems. The amount of DSP logic required by some protocols (*DSL) would make per-apartment boxes rather expensive compared to the installation of some additional physical wires. In rural areas, traditional telephone modems would still be rather common.

Mobile phones would be very popular. Their computational specs would be rather low, but most of them would still be able to access Internet services and run downloadable third-party applications. Neither of these requires a lot of power -- in fact, every microprocessor is designed to run custom code to begin with. Very few phones would have built-in cameras, however -- the development of cheap and tiny digital camera cells has a lot to do with Moore's law. Also, global digital divide would be greater -- there wouldn't be extremely cheap handsets available in poor countries.

It must be emphasized here that even though IC feature sizes would be in the "1996 level", we wouldn't be building devices from the familiar 1996 components. The designs would be far more advanced and logic-efficient. Hardware milestones would have been more like "reinventing the wheel" than accumulating as much intellectual property as possible on a single chip. RISC and Transputer architectures would have displaced X86-like CISCs a long time ago and perhaps even given way to ingenious inventions we can't even imagine.

Affordable 3D printers would be just around the corner, just like in the real world. Their developmental bottlenecks have more to do with the material printing process itself than anything Moorean. Similarly, the setbacks in the progress of virtual reality helmets have more to do with optics and head-tracking sensors than semiconductors.

2. People would be more conscious about the use of computing resources.

As mentioned before, digital storage would be far less abundant than in the real world. Online services would still have tight per-user disk quotas and many users would be willing to actually pay for more space. Even laypeople would have a rather good grasp about kilobytes and megabytes and would often put effort in choosing efficient storage formats. All computer users would need to regularly choose what is worth keeping and what isn't. Online privacy would generally be better, as it would be prohibitively expensive for service providers to neurotically keep the complete track record of every user.

As global Internet backbones would have considerably slower capacities than local and mid-range networks, users would actually care about where each server is geographically located. Decentralized systems such as IRC and Usenet would therefore never have given way to centralized services. Search engines would be technically more similar to YacY than Google, social media more similar to Diaspora than Facebook. Even the equivalent of Wikipedia would be a network of thousands of servers -- a centralized site would have ended up being killed by deletionists. Big businesses would be embracing this "peer-to-peer" world instead of expanding their own server farms.

In general, Internet culture would be more decentralized, ephemeral and realtime than in the real world. Live broadcasts would be more common than vlogs or podcasts. Much less data would be permanently stored, so people would have relatively small digital footprints. Big companies would have far less power over users.

Attitudes towards software development would be quite different, especially in regards to efficiency and optimization. In the real world, wasteful use of computational resources is systematically overlooked because "no one will notice the problem in the future anyway". As a result, we have incredibly powerful computers whose software still suffers from mainframe-era problems such as ridiculously high UI latencies. In a Slow-Moore world, such problems would have been solved a long time ago: after all, all you need is a good user-level control to how the operating system priorizes different pieces of code and data, and some will to use it.

Another problem in real-world software development is the accumulation of abstraction layers. Abstraction is often useful during development, as it speeds up the process and simplifies maintenance, but most of the resulting dependencies are a completely waste of resources in the final product. A lot of this waste could be eliminated automatically by the use of advanced static analysis and other methods. From the vast contrast between carefully size-optimized hobbyist hacks and bloated mainstream software we might guess that some mind-boggling optimization ratios could be reached. However, the use and development of such tools has been seriously lagging behind because of the attitude problems caused by Moore's law.

In a Slow-Moore world, the use of computing resources would be extremely efficient compared to current standards. This wouldn't mean that hand-coded assembly would be particularly common, however. Instead, we would have something like "hack libraries": huge collections of efficient solutions for various problems, from low-level to high-level, from specific to generic. All tamed, tested and proven in their respective parameter ranges. Software development tools would have intelligent pattern-matchers that would find efficient hacks from these libraries, bolt them together in optimal arrangements and even optimize the bolts away. Hobbyists and professionals alike would be competing in finding ever smarter hacks and algorithms to include in the "wisdombase", thus making all software incrementally more resource-efficient.

3. There would still be a gap between digital and "real" content.

Regardless of how efficently hardware resources are used, unbreakable limits always exist. In a Slow-Moore world, for instance, film photography would still be superior in quality to digital photography. Also, since the digital culture would be far more resource-conscious, large resolutions wouldn't even be desirable in purely digital contexts.

Spreading "memes" as bitmap images is a central piece of today's Internet culture. Even snippets of on-line discussions get spread as bitmapped screenshots. Wasteful, yes, but compatible and therefore tolerable. The Slow-Moore Internet would probably be much more compatible with low-bit formats such as plaintext or vector and character graphics.

Since the beginning of digital culture, there has been a desire to import content from "meatspace" into the digital world. At first, people did it in laborous ways: books were typed into text files, paintings and photographs were repainted with graphics editors, songs were covered with tracker programs. Later, automatic methods appeared: pictures could be scanned, songs could be recorded and compressed into MP3-like formats. However, it took some time before straight automatic imports could compete against skillful manual effort. In low resolutions, skillful pixel-pushing still makes a difference. Synthesized songs take a fraction of the space of an equivalent MP3 recording. Eventually, the difference diminished, and no one longer cared about it.

In a Slow-Moore world, the timeline of digital media would have been vastly different. A-priori-digital content would still have vast advantages over imported media. Artists looking for worldwide appreciation via the Internet would often choose to take the effort to learn born-digital methods instead of just digitizing their analog works. As a result, many traditional disciplines of computer art would have grown enormous. Demoscene and low-bit techniques such as procedural content generation and tracker-like synthesized music would be the mainstream norm in the Internet culture instead of anything "underground".

Small steps towards photorealism and higher fidelity would still be able to impress large audiences, as they would still notice the difference. However, in a resource-conscious online culture, there would also probably be a strong countercultural movement against "high-bit" -- a movement seeking to embrace the established "Internet esthetics" instead of letting it be taken over and marginalized by imports.

Record and film companies would definitely be suing people for importing, covering and spreading their copyrighted material. However, they would still be able to sell it in physical formats because of their superior quality. There would also be a class of snobs who hate all "computer art" and all the related esthetic while preferring "real, physical formats".

4. Conclusion

A Slow-Moore world would be somewhat "backwards" in some respects but far more sensible or even more advanced in others. As a demoscener with an ever-growing conflict against today's industry-standard attitudes, I would probably prefer to live with a more moderate level of Moorean inflation. However, a Netflix fan who likes high-quality digital photography and doesn't mind being in surveillance would probably choose otherwise.

The point in my thought experiment was to justify my view that the idea of a linear tech tree strongly tied to Moore's law is a banal oversimplification. There are many other dimensions that need to be noticed as well.

The alternative timeline may also be used as inspiration for real-world projects. I would definitely like to see whether an aggressively optimizing code generation tool based on "hack libraries" could be feasible. I would also like to see the advent of a mainstream operating system that doesn't suck.

Nevertheless: Down with Moore's law fetishism! It's time for a more mature technological vision!


Visa-Valtteri 'visy' Pimiä said...

Like I commented on IRC, you didn't really explore the divergence of mass server storage, especially large magnetic tape dumps being extremely more common in this type of world, probably facilitating the emergence of vast data banks running on magnetic tape that people would be able to access and stream data from, instead of downloading things locally.

Being the only logical low-cost solution for mass storage, large adoption of the medium would lead to an era where the cost of tape would be very, very low, because China and other current real world emerging markets could have found their industrial niche in producing gigantic tape storage solutions.

viznut said...

Magnetic tape seems to be following Moore's law as well. According to Wikipedia, we had 12-gigabyte tapes in 1996. But you're right, both home users and server people would probably be dumping their extra data on tapes, so the digital footprints would actually be somewhat larger. However, all the data about everyone wouldn't be readily accessible to anyone in random-access means (some kind of slow retrieve-and-combine requests would be needed) so I guess we might still be somewhat safer.

We would definitely have services that stream from tapes. They would probably be multicast in nature, as opposed to real world's random-access unicast. Services like Netflix could be based on a somewhat limited number of TV-like channels whose content is based on users' requests. The more you pay, the sooner you'll get to see the movie you requested. Just watching would be cheaper or even free, of course.

Anonymous said...

We desperately need Moore's law or better. Computing is still in the dark ages. We have seen nothing yet, we need massive raw power for the things that matter. Computing is not about editing documents, beign social, etc. Thinking we do not need growth is the worst kind of lack of imagination you can have.

Spencer said...

I'm curious about the advent of hardware hobbyists. I would suspect places like radio shack and best buy would have a catalog of simple discrete components necessary for expanding computer hardware. Thru hole/ socket based ram would be fairly common, and I would hazard that cooling systems would be far more complex than "stick a fan on it". Granted, there are liquid cooling systems out there, but to to an average consumer Liquid and hardware don't mix. The level of decline in print media would be nothing. Newspapers and magazines would still dominate how people prefer to get information, strictly because according on the assumption that it would be cheaper to print than buying storage space.

Interesting thought juice.

tomfilipino said...

tape? are you guys kidding?
colling system? vapor machinery?

im prety sure that memory storage ou computation process (mainly) cost something, and yes heat is created. But on the comment lines we still in XIX.

isnt no one expecting a typical first order transition in computer tecnology in lets say in 50yrs?

no lets just live inside The buble, shall we?

keanna said...

APTRON is a chief training institute giving Best Linux training in Gurgaon from over two decades. APTRON offer one of a kind learning knowledge with the best foundation and most recent instruments. The course educational plan is planned with the goal that the applicant can begin rehearsing as the expert Linux engineer when they complete their program.
For More Info:- Linux course in Gurgaon

independent said...

Oh, the article is very useful. I like it a lot. I also have similar articles. Click. ความงาม | สุขภาพ | โรคกะเพาะ | โรคผิวหนัง | ข้อดีของการนอนกอดแฟน

edok69 said...

This article is very good. We would like to leave a good article. To read more articles, click here >>>

No Name said...


* Please don't ask for sample or screenshot
* Payment in advance
* Time wasters or cheap questioners please stay away
*You can buy for your specific states too

Price $5 per each CC



Contact Us

-->Whatsapp > +923172721122
-->Email >
-->Telegram > @leadsupplier
-->ICQ > 752822040


First Name | Last Name | SSN | Dob | Address | State | City | Zip | Phone Number | Account Number | Bank NAME | DL Number | Home Owner | IP Address |MMN | Income

*Hope for the long term deal
*If anyone need leads In bulk, I'll definetly negotiate


UFA747 said...

Nice Information! I personally really appreciate your article. This is a great website. I will make sure that I stop back again. These are some really great tips! Another important note is to make sure you give completely specific instructions to your cleaning staffs.

Thanks,10 เว็บยอดนิยม

IAHIA said...

"ซิมิคาส บินด่วนถึง เมอร์ซี่ไซด์>> ข่าวฟุตบอล "

หวยเด็ดหวยดัง said...

I will be looking forward to your next post. Thank you
อัพเดทข่าวหวยตามกระแส >>> ทำนายฝัน ฝันว่าไปสวนสัตว์ คำทำนาย พร้อมเลขมงคล

No Name said...

Selling USA FRESH SSN Leads/Fullz, along with Driving License/ID Number with good connectivity.

**Price for One SSN lead 2$**

All SSN's are Tested & Verified. Fresh spammed data.



->Bulk order negotiable
->Hope for the long term business
->You can asked for specific states too

**Contact 24/7**

Whatsapp > +923172721122

Email >

Telegram > @leadsupplier

ICQ > 752822040

IAHIA said...

edok69 said...

Suggest good information in this message, click here.

No Name said...

Selling USA FRESH SPAMMED SSN Leads/Fullz, along with Driving License/ID Number with EXCELLENT connectivity.


>All Leads are Tested & Verified.
>Invalid info found, will be replaced.
>Serious buyers will be welcome & will give discounts to them.
>Fresh spammed data of USA Credit Bureau
>Good credit Scores, 700 minimum scores.

Email >
Telegram > @leadsupplier
ICQ > 752822040



->Bulk order will be preferable
->Minimum order 25 to 30 leads/fullz
->Hope for the long term business
->You can asked for specific states & zips
->You can demand for samples if you want to test
->Data will be given with in few mins after payment received

**Contact 24/7**

Email >
Telegram > @leadsupplier
ICQ > 752822040

chanee said...

Great article for you
ปั่นสล็อต ยังไงให้ได้เงิน
เว็บสล็อต เปิดใหม่ ล่าสุด

charles said...

Hello everyone I want to introduce you guys to a group a private investigators who can help you with information you need in any situation in life and they are ready to follow you step by step until your case is cleared just contact +17078685071 and you will happily ever after

Anonymous said...

Good to know about the information you have shared with us. I think your tips and strategies could help the writers to do their works properly

google cloud platform training in pune

pandith13 said...

Great post, thanks

shri chakram astro centre said...

it was great information and very useful

SRICHAKRAM ASTROLOGY.Best Astrologer In Marathahalli

shri chakram astro centre said...

This is a good article. Thanks for sharing


Vasudeva said...

This is a good article. Thanks for sharing

SRIKRISHANA ASTROLOGY.Best Astrologer In Shivamogga

pandit sri rajshekar bhat said...

I seriously love your site.. Very nice colors & theme.
visit here
Vashikaran Astrologer in Chamrajpet

Gokul Das said...

Its a really very impressive blog.Thank you for sharing.
Visit Best Astrologer in Nirmal

Arul Raj said...

Very good article,Thank you.
For sofa renovation contact thesofastore they gives best
Sofa Repair Services in Jayanagar

Sudharshan Bhat guruji said...

Good Article you know your article it's useful me. it's really good and helping your article and yes thanks for sharing information
For Astrological solutions contact Shri Durga astro center,They gives
Vashikaran Astrologer in Hubli

Expert Market Research said...

The global Office Supplies Market is estimated to grow to $247 billion by 2020. Items in the office supplies' market include paper-based products, storage and equipment, and writing and marketing instruments. It also covers non-core office supplies such as kitchen supplies, cleaning supplies, office furniture, workwear safety, computer consumables/accessories, etc. The global office solutions market can be divided into Recurring Spend products which are ordered on a regular basis (predominantly core office supplies and some non-core office supplies) such pens, paper files and folders, clips, staplers and markers and Non-recurring Spend Products.

โรซ่า said...

Hello, we have a game to recommend.


Thank you for your interest."

onlyblog said...

Nice Blog. Thanks for sharing with us. Such amazing information.

How to enhance the Images of Your Fashion Blog Effectively?

Anonymous said...

Hello Everyone

Selling SSN DOB Fullz in bulk
DL & work history will be include (if needed)
High Credit Scores USA Fullz/Pros

USA/UK/Canada CC Fullz also available
Dumps Tutorials & Dumps with pin codes available (101 & 202)
Business Fullz with EIN
Office365 Logs in bulk

Here I'm

TG/ICQ = @killhacks
WA = +92 317 2721122
Emails = exploit dot tools4u at gmail dot com
Wickr = peeterhacks

Hacking, Spamming, Carding Stuff also available
Tools with tutorials & Ebooks
Kali Linux Full course
D**P/D**k web Courses with Legit Links

Legit & Genuine Tools

Emails = exploit dot tools4u at gmail dot com
Wickr = peeterhacks
TG/ICQ = @killhacks
WA = +92 317 2721122Hello Everyone

Selling SSN DOB Fullz in bulk
DL & work history will be include (if needed)
High Credit Scores USA Fullz/Pros

USA/UK/Canada CC Fullz also available
Dumps Tutorials & Dumps with pin codes available (101 & 202)
Business Fullz with EIN
Office365 Logs in bulk

Here I'm

TG/ICQ = @killhacks
WA = +92 317 2721122
Emails = exploit dot tools4u at gmail dot com
Wickr = peeterhacks

Hacking, Spamming, Carding Stuff also available
Tools with tutorials & Ebooks
Kali Linux Full course
D**P/D**k web Courses with Legit Links

Legit & Genuine Tools

Emails = exploit dot tools4u at gmail dot com
Wickr = peeterhacks
TG/ICQ = @killhacks
WA = +92 317 2721122

vrindha said...

The browsers that have been affected by it will reduce performance and cause slow work. Users will automatically redirect to fake websites. Because of this, users become annoyed by these websites and need help using them effectively. The biggest issue with this type of threat is that it operates hidden in the background and gathers details like banking login information such as IP addresses, IDs, and credit card numbers, as well as passwords of social sites and other information.

rts dsrlte

안전놀이터 said...

success is always for you, yes!

카지노 said...

Hello friends, its great post about tutoring and completely defined,

토토 said...

keep it up all the time.

먹튀검증 said...

Great delivery. Outstanding arguments. said...

I will certainly digg it and personally suggest to my friends. said...

I am confident they will be benefited from this web site.

casinositeguidecom.JDS said...

I must thank you for the efforts you’ve put in writing this blog.

safetotositepro.JDS said...

I am hoping to view the same high-grade blog posts from you later on as well.