What I Think I Want for Maintaining Web Pages

I have heard similar comments elsewhere. I claim no originality. This is rather like what software developers need. There are differences that may be important.

Today I try to keep my complete set of pages both on my Mac at home and on the machine where I rent space and which actually serves the pages. Once an earlier service lost all of my pages and I was able to restore quickly from my Mac. Also there are searches that I can do on my Mac, with BBEdit, for which grep does not suffice. With:

I might not need up-to-date html files on my Mac.

One of the problems currently is that I spend more time futzing than writing. I must keep files up to date on both machines. I have occasionally copied the file in the wrong direction, losing updates. At least once I copied a ruined file on top of the only remaining good copy. Operator error for sure, but I make errors.

I think that the following is possible and actually not a great deal of work:

Another computer service under different management and a different geographical location than my web site, backs up my site. It makes chronological history of states of my site with perhaps one week resolution. The work required is only comparable with the work of supporting my editing changes in the first place. In other words next to nothing. Techniques for remote comparison reduce these costs to nearly zero. Backwards delta schemes make access to recent material nearly as cheap as a normal server while the distant past is only a bit more expensive. The perpetuity storage costs are orders of magnitude less than the cost of the original editing work. Accessing very old data costs a perceptible amount.

I also want facile difference service without having to transmit files home just to see if they might be different.

Software Development

Software developers need named snapshots of their software and also need trees of evolution. They need merge capabilities and they may need to have the data on their computer for more efficient access. Also they are less likely to find storage on a public server to provide adequate protection for intellectual property.

Business Models

Two models come to mind: In the first the backup service keeps the data in proprietary form and does not export the compressed updates. This potentially holds that data hostage. Presumably they charge less. In fact they might charge only for access. I have suspected that this might be Achieva’s plan for they claim to archive the web but have no guaranteed price for access to old data.

The second model is that the service makes the compressed archive data available in a public format. One CD could easily hold 200 weekly versions of my 450 page, 3MB web site.

Note that the remote comparison stuff above requires functionality not easily provided in Unix. See virtual.