Before Google was Google, it was BackRub, a Stanford research project with a crawler, a pile of scavenged hardware, and a name that probably would not have survived a marketing meeting.
Visit the surviving BackRub page
BackRub grew out of work by Larry Page and Sergey Brin at Stanford in the mid-1990s. The central idea was to use the Web’s link structure as information about importance instead of relying only on the words appearing inside a page.
That sounds ordinary now because the idea helped reshape search.
The old project page is interesting because it shows the system while it was still a research machine rather than a corporation.
Its rough statistics from August 29, 1996 reported about 75.23 million indexable HTML URLs and roughly 207 gigabytes of downloaded content. The page says BackRub was written in Java and Python and ran across Sun Ultra and Intel Pentium machines using Linux. The primary database sat on a Sun Ultra II with 28 GB of disk.
Twenty-eight gigabytes.
There are phones now that would consider that an insult.
Search while the Web was still small enough to describe
The page also credits Scott Hassan and Alan Steremberg for implementation help and Sergey Brin for major involvement. Larry Page signs the note at the bottom.
That kind of detail is why surviving project pages matter.
Modern company histories usually compress the origin story into a clean sequence: two graduate students, a clever ranking method, a garage, Google.
The BackRub page is messier and better. It talks about URLs attempted, pages not yet attempted, robots exclusions, socket errors, programming languages, hardware, and disk capacity. It reads like a system somebody was actively trying to keep running.
The National Science Foundation’s history of the project describes BackRub as a prototype built from Stanford’s Digital Library work. Page and Brin used links to construct a ranking method that developed into PageRank, and by 1998 the project had moved toward the company that became Google.
CacheRat’s 1,967 Ancient Web Domains research list includes pages like this because the old Web sometimes leaves the origin of an enormous modern system sitting in ordinary HTML beside statistics somebody typed by hand.
No anniversary redesign required.
Just the machine before it became an empire.