i should be working (but i'm fixing my computer).

Since my my kde4.2.3 (etc.) is running pretty well, my attention has turned again toward cloud computing solutions.

Follow up:

As I will describe later in this article (after a good deal of rambling) I am considering some php solutions to document production.

Over the years, there have been repeated predictions that we would give up our personal computers for a "web appliance". That always was under the spectre of some big corporation owning all of my hardware, software and data at some unaccessible remote location.

So, I was one of the most resistant to the idea. Plus, it seemed to me it would be "slow" -- at least in the foreseeable future.

Time has a way of eclipsing both problems and solutions. One example is archive storage -- stacks of floppies > tape drives > CDs > DVDs > giant hard drives > online storage. It seems every time I have a workable solution in one media I notice that while I have been solving the problem, a better solution has presented itself without me noticing.

So, I am probably in the process of solving a problem that will be irrelevant by the time I have a satisfactory solution of my own.

With internet speeds becoming adequate and ubiquitous and with net books and smart phones quickly merging to form highly portable web appliances, online solutions are becoming more appealing. But, even there, the paradigm is not obvious. Do we need to move the software onto the server (eg., everything works in the browser) or will the web appliance be small enough and powerful enough that we can still have substantial client side processing?

The thing I have noticed is that server-side solutions really reduce the time lost in setting up additional clients. It also allows instant propagation of changes -- change the system on the server and that's it. If I can access everything through an ordinary browser, then just about any off-the-shelf appliance is good to go. Not to mention the huge per-appliance-cost if we are talking about proprietary software solutions.

Whereas, my normal paradigm includes highly customized software installs. But, if all of this could be installed on something the size of flash drive -- and then duplicated from flash drive to flash drive (eg., Portable Linux), the client-side solution remain viable.

The problem there is that the average public computer location (eg., local library) is not going to let you reboot the machine into Portable Linux out of security concerns. Yet I perceive a future in which public locations have what is essentially a docking station -- a keyboard, large screen, mouse and printer setup -- into which you slap your smart cell phone.

So, achieving the correct balance of client/server tools is important. All that said, I have successfully moved a lot of formerly local applications to my hosted website and into my browser. My data applications are running through php. My documentation and note taking are browser-based. My email is IMAP. My phones are in the process of moving online (more on that in a later post).

There are two significant functions for which I haven't found satisfactory server side solutions: encryption and word processing.

Naturally, there are various word processing online solutions already (eg., Google Docs) -- but these solutions have all of the objections that prevented me from buying into the web appliance concept in the first place. That is, lack of control.

My hosted web server has been very, very reliable. It maintains professional backup. Also, I have the option of running my own web server if I so choose. (I'm uncertain about the security issues of the hosted account. Obviously, a good deal of information there is unencrypted and therefore accessible to employees of the server company. I encrypt data that needs to be secure to gpg keys. The problem with this is that over time as computers increase in power, encryption becomes less secure. If someone has access and adequate time, it's likely they can sooner or later decrypt anything.)

But, there is a trade-off between secrecy/paranoia and successful workflow. As far as I know, for the foreseeable future, my hosted data is adequately secure.

In any event, with a hosted account, I have substantial control over my own data. With PHP, Javascript and other programming solutions I have substantial control even over the software available on my site without having actual access to the root partition of the server.

So -- now about some document solutions.

First, I can't find any open source document production solution that is ready to use. We can easily create an html document and print it. But, HTML is not suited to printing needs.

There are some available efforts to translate Word or ODT documents to HTML and then back again. For instance, eyeOS developers provide an interface to OpenOffice for this purpose. But, apparently it worked with OpenOffice 2.0 and does not presently work with OpenOffice 3.0.

In addition, OpenOffice must be installed and accessible on the server, which precludes its use on hosted servers -- although it is apparently possible to redirect the document translation to another server.

There are also some online services which presumably can open Word Docs in a browser for you.

I haven't looked too closely at either of these solutions.

However, I'm looking at two avenues at the moment. One is, it should be possible to build a printable file using html table structure -- which can have a header and footer. But, one issues is that the table apparently won't page break in the middle of a row. So, how do you achieve text flow?

My thought is to write a php routine that parses the document, creates row breaks as needed, etc. Depending upon the level of sophistication needed for the final document, this should not be too difficult -- just really time consuming, mostly for debugging unforeseen problems.

But, there is another potential avenue as well. There are some HTML to PDF tools. For instance, I have a lot of data in a PHProjekt groupware system. PHProjekt has utilities for outputting certain reports to PDF.

While I have done some hacking on portions of the PHProjekt code to customize it for my own uses, I have never studied the PDF related code. But, this may be an option that will avoid re-inventing the wheel.

Also, some PHP code is available elsewhere for this purpose, for instance, here.

I don't care much about making my online documents doc or odt compatible. I just need to be able to produce a properly formatted printed document -- however is the most efficient way to get there.

Oh, the most efficient way to get there is if some computer major college student builds an open source program that meets my needs.

Which is partly why this post has gone into detail on the underlying system conceptualization. Should anyone read this (I know, not likely) it might help spur some thinking towards solutions I would like to see developed.

(The other goal would be to get gpg implemented into browsers in a way that we could slap a flash drive key into any computer and be able to retrieve our encrypted documents from the web browser. FireGPG does a nice job with the encryption/decryption, but is an "add-on" -- so not available to every computer -- and also requires that gpg and the relevant keys be available on the host computer. If there could be a "standard" allowing gpg encryption on any browser, with a transportable key, I'd be very happy. AND, BTW, not one of those "governmental" solutions where the feds get to have their own personal key ...)

The search continues ...