Thank you for all the responses and kind words. I'll try to answer the all questions posed in this thread below. I'm going to fix a few of the more obvious bugs, and incorporate some small improvements on Sunday, so there should be a new release early next week.
···
----
Tom Counsell. http://tom.counsell.org
On 11 Jan 2005, at 19:36, Florian Gross wrote:
Any demo wiki online? The automatic linking sounds interesting. I'd like to see that in action
Well, in my announcement I mentioned that Soks is slower than your average wiki and therefore not suited to on any 100MHz 64MB virtual hosts. Well, http://soks.counsell.org is on just such a cheap, highly contested, low powered virtual host. It may give you a bit of an impression, but it is slow, and if it may need to go off-line pretty fast. If anyone is willing to offer a higher performance space I'd appreciate that (although I understand that might be a bit of a leap of faith given it is only first release).
On 11 Jan 2005, at 19:46, gabriele renzi wrote:
Actually I think it's not only interesting but also scaring. I'm afraid of autolinking an "in" or "to" page 
A really frequent autolink would be annoying and would slow the wiki down. It could be straightforwardly solved by deleting the page titled 'to' or 'in' (which is easy to do). You are right to bring it up though, I clearly haven't thought properly about what a malicious (or clumsy) user might do. I'll put thought into it.
On 11 Jan 2005, at 21:51, Florian Gross wrote:
So how do I create a new page? I think the work flow in previous Wikis was link word, click word, enter content. Maybe this could use something like Ctrl-clicking a word (yeah, JavaScript, there also needs to be another way of doing this, but this one will be the more rapid one) to go to a non-existing page?
As Bill Atkins kindly mentioned, currently you click on a new page link. You can also, if you want to keep the usual wiki work flow, use bracketed wiki links. Your mention of JavaScript set me googling and turns out you can use JavaScript to enable you to highlight a word or phrase, press ctrl-n, and have the browser skip to editing a new page with the phrase as its title. So thanks for the brilliant inspiration! Will appear in the next version (however only certain browsers seem to do this Javascript in a way I can understand. On a mac Safari seems to work, but not Omniweb).
On 11 Jan 2005, at 22:11, Douglas Livingstone wrote:
Another thing I miss is the linkback, where you click the title of a
page, and you see a list of all the links to that page.
Link back is implemented. The links currently appear in the side-bar. This may be a little cramped and unintuitive for seasoned wiki users, so I'll make clicking the title give a link back in the next version.
How far are you planning to take this wiki Thomas?
Not sure yet. I'm definitely keen on the concept of wikis, thing there are lots of improvements to be made, and have a few evenings a week free to work on this sort of thing for at least the next two years (doing a PhD at present). I'll definitely try and stabilise this version and get the RSS feeds working. I'd also like to write some import/export filters to different wiki formats. After that I'm wondering is whether I should spend my time helping to put some of my ideas into instiki or ruwiki rather than continuing to develop a third way.
I'm also wondering about working on _why's RedCloth, which is great, but the errors I occasionally get suggest its enormous and fiendishly cunning regexps are pushing String#gsub to the limit and I was wondering about rewriting it using a StringScanner and doing the conversion on a token by token basis.
On 12 Jan 2005, at 03:15, James Britt wrote:
Sweet! I just installed it on a Win2K box.
Chuckle... "With Soks, there really *is* no step three!" (TM)
Time to play around ...
Did it work?
A suggestion: Make the pages use XHTML transitional, or at least HTML 4.01, with a proper doctype and all that w3c stuff.
Fair point. I'll put all the doctype stuff in and do what I can to make it validate. Don't know if I can guarantee validation on all pages, what with all the messing with the textile to html conversion. Would be a good goal.
On 12 Jan 2005, at 05:51, Alexey Verkhovsky wrote:
If you refer to a certain Instiki bug (not creating storage/2500 directory on startup), it has been fixed in release 0.9.2
I like Instiki, and experimented at one point with editing its code to make it do what I want, but found it heavy going. I understand that Instiki is going through a major re-write now. I hope that will make things easier for me to understand, and then I can perhaps help merge some of the Soks ideas into Instiki. (not that I'm saying the Soks code is any easier to understand, just easier for me to understand).
On 12 Jan 2005, at 14:29, Ruth A. Kramer wrote (To Alex, but I'd like to answer as well):
The first bug, IMHO, is that HTML entities don't work as I'm used to,
they are not saved as HTML entities but instead are converted to the
"real" character, thus, for example, if you (I) want to display
something like <an example HTML tag>, it disappears after (iirc) the
next edit.
I haven't noticed this in instiki? perhaps it depends on the tag? A word of warning. At the moment Soks does not filter any tags which is clearly not a good thing if you suspect malicious users. _why has just published a handy piece of filter code on his blog which I'll use on the next version.
On 12 Jan 2005, at 15:02, Andreas Semt wrote:
Does exists a converter from Textile -> PDF? Would be great! So i could print the wiki content to give it to some people who "love" printed pages 
I know that instiki has / had this. I think it converted textile into latex and then latex into pdf and was therefore dependent on having the appropriate system libraries. It is an interesting idea, and probably of particular use to projects like http://www.hieraki.org/\. I notice that Austin Ziegler started a pure ruby pdf writing project on rubyforge http://rubyforge.org/projects/ruby-pdf/ but I don't know if that has developed far enough to be a useful starting point.
