OpenPA and Internet History

When OpenPA was founded in 1999, companies like HP just started to open up to open source projects, the Internet and the public interest in general and began publishing much information freely on the web. The main job for OpenPA at the time was to find, analyze and correlate this growing heap of distributed information into a single, coherent resource on PA-RISC – which was missing at the time.

PA-RISC computers were stilly widely used and just about to be resold to second hand users, open source projects – forming a growing, active public community.

The main challenges for OpenPA at the time were both finding all the available information, as search engines were still young in the late 1990s, as well as making sense of it all as it was just so much and new sources kept appearing. This went on until the mid to late 2000s, when solid and stable sources could be found and referenced, which OpenPA did.

The Internet and information on it changed since then, slowly but surely, in a profound way. Many original sources have disappeared and so much information has been lost in only two decades – making OpenPA the authoritative source for PA-RISC in some ways. A long journey from documenting complex information of the 1990s to an historic archive on the PA-RISC era.

Over two decades of work researching PA-RISC and HP 9000 resulted in over 100 articles on OpenPA (now 119) reflecting the broad PA-RISC landscape and Internet periods passed.

The Internet never forgets

It seems the Internet does forget.

When much of PA-RISC information and many sources became public in the late 1990s, there was no real reason to believe these sources would soon disappear any time soon again. It seemed most would find a good home in repositories, search engines, archives and hobbyists sites (such as GeoCities et al).

However, that changed only a few years later.

OpenPA started right in the period around the dot-com bubble where much documentation and IT was easily available and the Internet and content on it looked like it will be around forever.

During the economic and corporate upheavals of the following decade lots of information (not only PA-RISC) fell by the wayside and the nature of the Internet and how companies used it changed. Journals closed or got rid of their archives, websites plainly vanished, companies merged and removed or lost old documentation in the process. Few primary sources remain.

The Professional Period

HP generally produced a great deal of documentation and specific information on PA-RISC over the years on various channels. From PA-RISC inception in the 1980s until the late 1990s, much of that information was confined to commercial, industrial and research relationships and not open to the general public. Partners and clients had access to a great deal of documentation on PA-RISC hardware, architecture and software.

A complete ecosystem was available around HP products including HP 9000 and PA-RISC – user groups, conferences, trade journals and specific channel distribution. HP, the HP 9000 series and HP-UX were focused strongly on industrial, engineering and instrumentation customer bases, so documentation and PA-RISC information tended to stay confined to those circle, with less of academic (public) exposure than other vendors and architectures.

It was hard to get at HP and PA-RISC documentation at those times if not a partner or client. Even though called Information Age already, access to it was tightly confined in the 80s and 90s to those with vendor links or deeper technological interests. This was probably one of the reasons HP 9000 and HPPA computers were not documented as well as other architectures.

It was all very conservative and professional until the the end of the Nineties, the era OpenPA started in.

The Open Era

Around the time open Unix architectures and specifically Linux took off in the late 1990s, HP began to take notice of fledging open source projects for PA-RISC with potentially wider distribution. Many PA-RISC computers found their way to larger customer bases including university research projects and development groups.

This created more demand for proper documentation on PA-RISC systems outside the usual sales and vendor channels. Lots of information on PA-RISC was released that way between the late 1990s and early 2000s, plus many business magazines and industry journals started publishing on the web, including access to their archives.

There was almost too much information available at the time. Search engines were still new-ish and often still directory-based and many public and specialized repositories were available. In this open period, a splurge of documentation was released by vendors from their commercial, sales and research archives and put on the web freely with no questions asked.

You could even get printed architecture books for free from some vendors – like Motorola, and some operating system releases were made available due to the Y2K scare with few strings.

It was a great time researching and documenting PA-RISC, and more and more repositories and sources kept appearing. Finding and sorting all this information from expanding sources proved maybe the hardest, as there was so much to choose from and OpenPA took its time.

Fleeting Information

But PA-RISC documentation and information changed after these few years of open access.

It slowly started in the mid-2000s with corporate upheavals and the technology landscape changing profoundly. Linux and open source projects began arriving at scale in many formerly traditional Unix niches and the Intel plus Windows (NT) ecosystem took swathes out of the RISC workstation market.

RISC and Unix lost significance in the market, for clients and thus for vendors. IT had to become cheap and easy, a train which the conservative Unix servers and workstations did not really catch. These traditional vendors slowly lost interest in RISC computers and their business units too and providing documentation on increasingly unloved products became an afterthought.

The financial crisis at the end of the 2000s led to more corporate upheavals and economic hardship and again more IT consolidation followed. Departments closed and took their websites and documentation with them. Whole business units disappeared with their products. The business press and trade journals were hit also, and many publishing repositories with articles from the exciting 80s to 2000s information technology vanished without a trace or archive.

Many original PA-RISC sources and documents vanished during that time of the 2010s, as the commercial RISC and Unix era got slowly forgotten and information on it too.

Much of the later content of OpenPA thus had to make more use of secondary sources and industry articles that remained available – which in turn also started disappearing from the web during the 2020s. There was not much left, compared to early 2000s – gone were most original HP documentation on PA-RISC, contemporary news articles, news releases, prices and even ads.

RISC History

It got quite difficult to maintain articles based on ever fluctuating sources. This might just be the transitional nature of the Internet, but it was surprising to see so much go after doing this for over two decades. And it got worse in the meantime – so much has been lost as of the 2020s, both official company archives and secondary sources like the press.

This site has become an historic archive on PA-RISC and HP 9000 now since many other sources are just not around anymore. Originally envisaged to be a platform to filter and consolidate all the information the new Internet started to offer on HP RISC, in the span of two decades OpenPA became the sole source of information for many aspects of PA-RISC and HP 9000.

When OpenPA started, HP 9000 machines were still (quite) current and PA-RISC a somewhat actively developed and marketed architecture. Both PA-RISC computers and documentation on them have become historic since, and much of the information and many anecdotes were (or are) at risk of disappearing, unless documented somewhere.

Who still knows that PA-RISC ran not only NeXTSTEP, but also Windows NT and Netware? The PA-RISC 1.0 processors (and computers!) of the 80s were almost forgotten by the time OpenPA started in 1999 – a whole chapter of PA-RISC had no real product names

OpenPA will try to document and archive as much of the information on the PA-RISC part in the colorful era of RISC and Unix workstations as possible.

↑ up

Sources

Information on PA-RISC and HP 9000 used by OpenPA is based on public sources from public repositories. Primary sources were mostly HP and PA-RISC technical reference manuals, handbooks and architecture guides from the 1990s that were provided publically by HP in the early 2000s plus product brochures added later.

This knowledge base on OpenPA was extended during the early years with secondary sources such as magazine articles, news releases and journals like the HP Journal and the professional business press. It has been a long journey and moved from documenting then current Unix PA-RISC workstations to an almost historical archive on the PA-RISC side of the 1990s Unix era.

PA-RISC CPU, chipset and architecture: These articles were mostly based on primary sources from HP like the great HP External Reference Specifications (ERS) documents and technical publications from HP CPU and chip design labs made available during the 2000s for HP-supported open source projects. Secondary sources in the form of articles or academic papers were used as well.

Computer systems: Information on the HP 9000 and PA-RISC computers is based on primary sources from HP and vendors in the form of system user guides, technical handbooks and architecture white papers. Also used were marketing brochures, news articles and industry reporting.

Operating systems: The pages are either based on official user and admin documentation as well as academic papers, talks, whitepapers and such. The heydays of open source, Mach, Linux and BSD research systems from the 1990s resulted in much information on public websites. As much of the operating system development on PA-RISC happened during 1990s and then early 2000s, information is getting sparse now.

Paper books: Due to the effects on Internet sources described above, alternatives were searched. A large body of literature exists in academic papers and R&D conferences of the 1980s and 90s, when HP published frequently on the new Precision Architecture and RISC computers. CompCon digests of papers contain many interesting articles – mostly deeply buried in libraries. Some museal sites also have been digitizing many old product brochures from HP and others, another valuable resource.

↑ up

Closing remarks

For a taste of the 90s Internet, there is a fabulous archive of GeoCities on tumblr. Much HP 9000 information can be found at the HP Computer Museum as well as the Hewlett Packard section of archived brochures at 1000 BiT. The great HP-UX Workstation PA-RISC Hardware Compatibility List (mirror) from HP played a big role in the early years. Many business and tech journals are now available at the Internet Archive.