Dries Buytaert, co-founder of Acquia and an original creator/project leader for Drupal development, recently posted a blog about what he calls “The Big Reverse of the Web.” The basic idea is that the World Wide Web is moving from its original “pull” model towards a new generation of web service that is based on a “push” model.
There are many examples of how we traditionally pull information. For instance, when you buy a newspaper (electronic or not) you receive a “package” of information that includes current news, opinions, advertising, classified ads, etc. You then select what you are interested in actually reading. It could be argued, however, that most of the newspaper was “pushed” to (or at) you because you didn’t ask for all of it – you probably didn’t ask for the obituaries.
Traditional websites require the user to first find the content, either by knowing a URL or by using a search engine, and then to retrieve the contents. Your search can be can be typed or spoken and be as broad or narrow as you want.
There are many examples of the pull model:
- You visit an app store to find, download (i.e., pull) and install a desired application;
- You request current weather information for your area (or any other location);
- You look up Wikipedia (or any similar site) and information is delivered as a web page; or
- You access your bank account (you need to know the account number) to view your balances.
The user provides most of the smarts and each transaction is treated as being essentially independent. Enhancements and shortcuts are possible through localization, personalization and history data, but selecting the desired content is still up to the user.
Push-based systems, on the other hand, operate differently – content is detected and delivered at least semi-automatically via subscriptions or preferences. Examples of push systems include:
- You sign up for automatic patching from Microsoft, which detects the O/S version you are using and delivers the correct patches (i.e., patch Tuesday);
- You subscribe to a news feed which delivers items for your stated areas of interest (e.g., Google Alerts);
- You receive weekly grocery flyers and coupons from a grocery store you frequent (e.g. reebee);
- You subscribe to music that is streamed from a radio station or music service (e.g., Apple Music); or
- You can get a wide variety of notifications on your smartphone.
To quote from Dries’ blog (referenced earlier):
Ten years from now we’re going to look back and recognize that search-based content discovery was broken. Today the burden is put on the user to find relevant content either via directly typing in a URL or by crafting complex search queries. While pull-based experiences might not go away; push-based experiences will dominate as they will prove to be much more efficient.
Many of you won’t like it (at first), but push will win over pull. Healthcare is going through a similar transformation from pull to push; instead of going to a doctor, we’ll have web-enabled hardware and software that is able to self-diagnose. Wearables like activity trackers are just the start of decades of innovation and opportunity in healthcare. Helped by the web, education is also moving from pull to push. Why go to a classroom when personalized training can come to you?
We are at the beginning of a transition bridging two distinctly different types of economies. First, a “push economy” that tries to anticipate consumer demand, creates standardized or generic products in large amounts, and “pushes” them into the market via global distribution channels and marketing. Now, a “pull economy” that—rather than creating standardized products—will create highly customized products and services produced on-demand and delivered to consumers through one-on-one relationships and truly personal experiences.
Transformation or expansion?
I wonder if this is really a replacement for the current web or if it’s really an expansion and maturing of the options. Changes such as these are often driven by embedding more intelligence into the service.
For example, if the push web knows you live in Toronto and that you frequently go to specific types and qualities of Japanese restaurant, then it can make better suggestions when you are travelling.
Another example – if you identify your interest in gourmet coffee (perhaps through Swarm check-ins), then instead of having to scan the general food news you could be notified whenever a new coffee shop opens within a mile of your home.
The first generation web has made an enormous amount of content available – web pages, documents, videos, software, maps, etc. It is almost essential for every business to have a web presence. Discovery assistance is available using Google or Bing but remembering to check for something new is not very user-friendly. For example, I frequently search for new Canadian cloud computing case studies – they don’t appear every day!
Today, the user has to figure out the search terms and filter through thousands of results to find what they want. We’ve all scanned through dozens of search pages to find something, often just to discover it’s behind a paywall.
The next generation of push web services will need to be more helpful! For example: “Hey Siri, let me know when a new coffee shop opens in my neighbourhood”. You will be notified If something new appears online. With context, your cloud-based “smart assistant” can provide a much more effective and efficient notification service.
The range of communications models
In his blog post Dries focussed on two specific models for interaction – two-party push and pull – but these are not the only ways to communicate.
Various classifications for communication styles exist, including:
- Number of communicating parties: Communications can be one-to-one (e.g., a telephone call), one-to-multiple (broadcast, e.g., a television channel), multiple-to-multiple (e.g., a conference call), or multiple-to-one (e.g., sensors or devices). In aggregate, many services will appear to be many-to-many.
- Single or multiple media: Communications can be single media (voice, text, data, photographs, etc.) or multimedia (such as emails with embedded pictures and graphics).
- One-way or two-way: Interactions could be uni-directional (e.g., television broadcast) or bi-directional (e.g., a telephone call).
- Realtime vs. store-and-forward: While all transmission paths have some delay, if the delay is minimal it is classified as realtime (such as with a telephone or radio). Email is an example of store-and-forward non-realtime communications.
- Connection-mode vs. connectionless-mode: The interaction may be pre-arranged (such as a dial telephone call) or may be best efforts (a text message, for example, can be sent with no guarantee it will be delivered or read immediately).
- End user (application) vs. system communications: A considerable amount of communications occurs “under the table” to control the process of communicating and to manage systems. For example, domain name services (DNS), time services, network management (SNMP), and others are not directly accessible to the end user.
- Inter-personal vs. inter-machine communications: One or more of the communicating parties may be a machine, serving either as an end user or as an intermediary. Person-to-person (P2P), person-to-machine (P2M) and machine-to-machine (M2M) are different categories.
The Internet, which has become the foundation for all electronic communications, is evolving from its simple roots into a web of intelligent services that enhance the human experience – the big reverse of the web is a major but not the only step towards that vision.
This is what I think – do you agree?