...
It is important to note that the AI Lab and LCS still occupy the same physical building in 1995, so a reference to ``the Lab'' means one of the two organizations, while ``the lab'' means the physical building.
...
We may eventually make an effort to preserve these tapes, but they have a relatively low priority.
...
DAT was chosen quite arbitrarily; it would be trivial for us to switch to any other medium at any time. TCFS is completely insensitive to the media and system it is written on. The cost and availability of DAT were the prime factors in our decision, not any technical merits of the medium.
...
This is the mechanism used for including the raw and translated directory listings in the ITS example.
...
One day, Dr. Thomas Knight may take on the task of assembling the necessary hardware to read these 7-track tapes.
...
In much the same way that librarians used to write notes on cards in the card catalogs of days gone by.
...
Variable names may be included separately. They may be useful, but they may produce too many false leads and bloat the concordance significantly. This issue has not been studied in enough detail to make a final decision.
...
Note that scanning someone's mail files introduces some privacy issues and we wouldn't permit such a scenario without the owner's permission.
...
Ideally, we would want a media jukebox that could give us on-line access to all the data 24 hours a day. At present, an auto-changer system able to contain the 262 gigabytes of ``important'' data (from Table gif) would be prohibitively expensive.
...
Furthermore, there is nothing on our backup tapes that would lead us to believe that they are an accurate representation of the file system. Our only assurance on that front is the paper dump logs.

boogles@martigny.ai.mit.edu