A Eulogy For AltaVista, The Google Of Its Time | Hacker News Reader

A Eulogy For AltaVista, The Google Of Its Time

searchengineland.com

103 pointsgregosaur49 comments

I remember when google came out I didn't quite understand what was different, and hated the fact that I couldn't use operators such as +-(). The feeling lucky button further reinforced the feeling that google was a dumbed down interface for the masses that wasn't for professionals. Took me a bit to realize that the results where better then AltaVista regardless of the reduced sculptability.

EDIT: haha now it's been so long I've even forgotten the operators. They where the boolean operators AND, OR, NOT etc. spelled out, and other very useful sculpting tools.

http://www.netstrider.com/search/altavista.html

In the early times of Google I switched back and forth to Altavista because the NEAR operator was the way to adjust the precision of my terms.

I remember when Google came out and vividly remember what was different. It was the UI. It was clean, sparse and without any ads.

For me the search results were across the board worse than Altavista but better specifically in searching API documentation.

It's hard to believe today but one innovation in the google home page was that you didn't actually have to click into the text box to begin typing. For some reason every other search engine at the time required you to do this one little extra step.

AKA the "Billion Dollar Javascript" - http://www.branded3.com/blogs/billion-dollar-javascript/

Thanks for the link, it was a great read.

In the initial time where alta vista were seriously co-existing as search engines (on that graph, in the time before they crossed paths in usage) Alta Vista was a little nicer because it really was a much more raw look at the web. If I did a search for 4 keywords in a row, the first ten hits were generally an exact match. (there weren't as many cloned sites etc... so this was a reasonable thing to do)

I miss that aspect of the web.

Really? I started using Google before it had its own domain, and was eagerly showing to people. AltaVista had started shitting up its interface with intrusive ads by then IIRC, but more importantly while the arcane combinations I needed to get decent results made me feel like a wizard, I could get as good if not better responses just typing a few terms into Google.

That's what I remember, but yknow, it's been over a decade so memories fade. I remember that it took me a little bit of time to transition from trying to guess combinations of words that would appear in my target documents to just utilizing keywords.

I miss altavista every single day.

Every time I search for a collection of terms, and then click the links, and page-search for one of those terms, and it's not in the page.

Google wastes my time like this 3-4 times per day.

You can use "intext:blah" to make sure "blah" actually occurs in the page body itself: http://www.googleguide.com/advanced_operators_reference.html

Just out of curiosity why is search text actually occurring in the page body an 'advanced' feature?

Because there are many factors taken into consideration by default, like anchor text of links to a page or the URL.

one factor is "search for keywords that the user didn't search for"

Well I ment that a little facetiously. However, I agree those things should be taken into account--in addition to including documents with the terms you ask for in them. Pages with my terms (in order) should be the absolute highest ranked, and it should go down from there. At least if we're talking about building an intuitive google vs. really aggravating google.

In real life, do you find something by mimicking it? Then why would web searching be any different?

Yes. "To find an X, you have to think like an X."

Because heuristic search is often useful. For example, I was recently running some searches for Final Fantasy 10, and was highly appreciative of Google also finding pages where 10 was written as X.

Of course, heuristic features in computer programs usually benefit from a "do what I say, not what I mean" override option, but it's reasonable to call that an advanced feature.

The problem is that Google frequently gets it wrong. A typical example: I was searching for a party's view of "wind turbines", and the second or third result was an article about that party's political "power" struggles. Bolded. Because wind turbine = power.

There's a fine line between clever and stupid.

This feature will probably be removed at the next Google search "upgrade". Just like allinanchor: OR allintext: OR allinurl: OR allintitle:

Sometimes it helps to use the Search tools | Verbatim option to keep Google from drawing unwarranted conclusions about a page's relevance.

They used to allow you to mark required terms with a +-sign, but unfortunately they took that feature out a while back.

Today you have to use quotes: "searchterm"

Pretty annoying as it requires two signs and at more than one place but at least it still allows me to search.

I miss Google from ~5 years ago

Me too, specially now with all the URL rewriting, language and IP filtering they do.

As I travel around the world, I need to keep changing my search settings to find what I want.

The URL rewriting is absolute bullshit. What is the deal with that?

Not sure if you are agreeing or disagreeing with me, my main issue is copy-paste from Google results, that now get mangled with their URL rewriting for SEO count.

Agreeing with you 100%.

Use http://www.google.com/ncr for "no country redirect". It's a life saver :)

Maybe licensing issues would prevent it, or make it difficult or costly to do, but I think it would be superb if at least some of the source code powering AltaVista were to be released under a liberal open source license. This is especially true if it's no longer a feasible product.

The code from the Digital era, for instance, would be quite a contribution to the Internet community, even if just for the sake of preserving historically-important software.

AltaVista has been a front-end for Yahoo search for >10 years... Not a lot of source code left...

Is this just speculation? Or do you have first-hand experience at Yahoo!, or something to that effect, to back up this claim?

No experience at Yahoo! here, but a quick look at the AltaVista page suggests the parent is correct.

The tabs eg. images, video, lead directly to subdomains of search.yahoo.com, so everything except web is handled very transparently on Yahoo's own pages.

More importantly, the action on the Alta Vista search form itself is (for me, I assume some of this is dynamically generated): http://us.yhs4.search.yahoo.com/r/_ylt=A0oG7pc8zNBRgwMAFJel8...

Note that the action on the yahoo.com (US) search form is: http://search.yahoo.com/search;_ylt=AkYpzVmc8WkqPKbMSU46mpqb...

It's certainly not complete proof, but it does seem that altavista.com search is handled by search.yahoo.com in a similar way to yahoo.com itself.

I'm not disputing that there may be very little to modern AltaVista. But that's not really what I was originally referring to, either.

I'm not sure about Yahoo!'s policies, but most other large companies or organizations with custom software have processes in place to ensure the archival of old source code.

In these kinds of organizations, it's quite typical to have source code going back many years, if not many decades, for each release of a given software product.

This has gotten even easier as source control has become more widely used, especially during the 1980s and later, and as storage space has gotten far cheaper and larger. Many development groups have put great effort into ensuring that entire histories have been preserved when moving between version control systems, even when several such transitions have taken place. It can sometimes be possible to see change sets committed back in the 1980s, for instance.

Again, while I don't have any first-hand experience there, I'd be very surprised if Yahoo! did not have at least some of the early AltaVista source code archived in some way.

It must be depressing sometimes to think of things like AltaVista that Yahoo had run into the ground under leadership that apparently didn't understand the assets they owned.

When AltaVista first appeared, I had used Lexis quite a bit, and had built my own inverted index for law forms. I thought "Wow, you can do that to Web pages!"

But it is even more depressing to keeping dragging the corpse of AltaVista around. Time to let go.

Exactly. Once Google beat it, no need to keep the corpse around. Not sure what Yahoo could do. Their own search had already surrendered.

I guess they had a significant amount of people who had never moved on. Better to serve them up something they knew than push they to Yahoo search and have then actually think about searching options.

Over time I guess that number become small enough that maintaining a separate brand is no longer worth the likely small amount of effort to do it.

Honestly, it's more depressing to think of the things Digital ran into the ground:

VMS (let's keep it expensive and hard to get ahold of while Unix is getting cheaper, and oh, then we'll decide we're a Windows shop anyway).

Flogging off their database technology to Oracle.

Alpha. Poor, poor Alpha.

The waste of giving the StrongARM to Intel, who would ultimately kill it.

AltaVista, of course.

Tru64's good bits (like AdvFS) was criminally wasted, although I guess that's more HP's fault in the end.

Very true. DEC could have been a major router maker, could have been Google, could have been a major CPU maker, could have been a major database vendor, and could have had a major piece of the PC business. That's got to be the world championship of neglecting great engineering (maybe excepting Xerox). Poor AltaVista went from a neglectful parent to a neglectful foster home.

I agree with Danny it is a sad sendoff for what was such a transformative technology. DEC never did understand open anything, and the Internet was always more of threat to them than a way to move forward.

babelfish.altavista.com was also the translate.google.com of its time

Ahhh babelfish. So many poorly translated high school spanish papers.

I remember the dark side, astalavista. My entry into the world of exploratory computing. ;)

Altavista's creator and original implementor was Paul Flaherty (1964-2006) (http://en.wikipedia.org/wiki/Paul_Flaherty). Altavista (altavista.dec.com) began life in his office at the DEC Network Research Lab in Palo Alto, CA. As Altavista usage grew, it consumed machines (alphas in those days), filling his office and moving into the hallway. Eventually it moved to its own facility and became something more than a research project.

Altavista was, pre-Google, the best of the search engines. I am fairly sure that the original idea for Altavista was a droll comment made by Andy Freeman at lunch that sketched the possibility of a smarter-than-inverted-list search engine for the Internet; Paul made it all happen.

I remember sitting around my friends house and he showed me AltaVista, I think previously I had been using Yahoo. He said have a go, so I typed some stuff into the box and then it listed a bunch of stuff, "cool eh?", my friend said.

"Well kind of", I replied, but none of the things it returned are what I'm looking for.

Oh no, that's not how it works, he explained. You have then find your actual results somewhere on the 2nd or 3rd page. Or narrow the results with a curious combinations of + and -.

If only I had connected the dots perhaps I would have written Google myself.

I remember teaching a friend how to research a topic by sitting her down in front of the computer, and showing her how to search all 5 major engines instead of just one, thereby getting a number of perspectives on the topic.

This article makes me seriously nostalgic.

I believe people like Jeff Dean and Sanjay Ghemawat who became instrumental in Google's success came from DEC. I'm sure they brought lot of talented folks from DEC that paved what Google's search dominance.

I remember the domain switch, and how previously the guy who actually had altavista.com used to have a little box on his personal website telling people looking for that other altavista search thing to go over here...

If my memory serves me AltaVista had good newsgroup support too. It was even possible to manually uudecode binary files from the search results.

Sadly, google was just easier to type.

Reply on news.ycombinator.com