Battle of the Boxes and Networks Summary

Battle of the Boxes

Hollerith's punch card system for the 1890 U.S. Census significantly reduced tabulation time compared to the 1880 Census. Hollerith's Tabulating Machine Company, later IBM, leased equipment and controlled client computing processes for optimal outcomes.

IBM's dominance in business computing involved leasing mainframes with included hardware, software, maintenance, and training. Unbundling in response to antitrust pressure allowed separate purchase of computers and software, fostering in-house programming and third-party software.

Mainframe computers and information appliances (e.g., Friden Flexowriter) were mainstream computing devices before unbundling. PCs from the late 1970s combined ownership with the flexibility of mainframes.

Separating software from hardware enabled computers to perform various tasks without mechanical modifications. PC word processing, spreadsheets, and databases improved upon older methods, while new applications like video games emerged.

PCs, initially for hobbyists, gained business traction due to their flexibility and affordability. Microsoft's vision of "a computer on every desk" promoted widespread software development. The Hollerith model contrasts with the appliance model; PCs provided adaptability and accessibility.

The PC's capacity to connect to the Internet, combined with user-maintained flexibility and code from various sources, enhanced its utility. This was supported by the development of plug-in architectures that allow for fourth party code on top of third party code.

A generative platform invites contributions, initially from amateurs, which is then refined and deployed by the market. The story of the PC versus information appliance reflects a recurring pattern where generative platforms lead to innovation, followed by challenges to openness.

Battle of the Networks

PCs succeeded due to their generativity which refers to open reprogrammability, supported by distributed ownership that removed barriers to tinkering. An alternative scenario would involve specialized information appliances, but networking is still essential for exchanging data.

Networking stand-alone appliances presents challenges, potentially converging on the Hollerith model. The Internet's generativity depends on usage cost, measurement, and user connectivity. Different network configurations affect levels of generativity, regulability, and control.

Utilizing the telephone network for data transfer faced regulatory hurdles. AT&T's control over its network and attached devices was challenged by innovations like the Hush-A-Phone and Carterfone, leading to third-party contributions.

Dial-up modems facilitated connections between PCs and computer networks. The Carterfone decision enabled third-party advances, transitioning the physical layer into a generative one that expanded activity in higher layers.

The first online services extending the IBM model include CompuServe and AOL that restricted subscribers' access to content and services deployed solely by the network providers themselves, which required network approval.

CompuServe depended on phone network generativity but lacked third-party tinkering. The proprietary service economic model was based on connect time charges, possibly restricting user contributions due to lack of formal relationships and risk concerns.

Early 1990s networks were corporate-run and non-interconnected with proprietary blends of offerings subject to content control. Mature firms favored exploitative over exploratory innovation by controlling user's abilities.

Before the Internet, PC owners used BBS (bulletin board system) software to create information services, depending on callers for content. Limited incoming lines restricted usage and PC generativity addressed limitations with software like FIDOnet.

FIDOnet relayed messages across numerous BBSs with associated PCs via the phone systems; however the network did not scale well. Proprietary services were better suited for mainstream consumers due to their reliability and ease of use.

The Internet, dating back to 1969, aimed to connect anyone on the network to anyone else, differing from proprietary networks. Its architects were academic researchers who sought government research funding to test protocols.

The Internet offered no specific information or services, it enabled connections. Its design facilitated cobbling together existing networks to make use as possible. This allowed the public to begin using the Internet which opened the door for a broad audience.

The Internet's design allows for open use, but user identification has been a problem because of anonymity. The lack of guaranteed transmission speed allows for potential network and abuse.

Peter Tattam created Trumpet Winsock, connecting Windows PCs to the Internet. Proprietary networks tried to design content and vet, but have been overwhelmed. Microsoft integrated this functionality into Windows 95, marking a move toward broader Internet accessibility.

Generative devices like PCs promote innovation, while less generative devices offer smoother operation. This creates trade-offs between proprietary and Internet services. The Internet frames made simplicity a goal based on trust.

The procrastination principle is the idea the network should not be designed to do anything that can be taken care of by its users. Another assumption is users are to be trusted, but has a lack of security concerns. Design decisions can impact everything from social interactions, potential abuses, and regulation.

The internet lacks "quality of service" which inhibits bandwidth guarantee. Therefore, there has been a debate about ISPs being able to prioritize certain data streams. Ultimately, the internet was designed for the research environment with the mindset there would be collaboration and cooperation.

The Internet experienced growth as well as expansion into the general public because the internet architecture was simple an fluid and backwater.