Raible Designs

Roller's Search Feature

I'd love to use Lucene to develop Roller's search feature, but there have been discussions of a much simpler search (i.e. simple SQL LIKE `%searchTerm%' statements). Since I'd like to dig in and learn Lucene, I think it would be valuable to integrate into Roller, but so as to not discourage any other ideas - maybe we could code for a pluggable search module. That way, roller administrators could choose which one they liked, and developers could plug in others.

I don't know how much time I'll have to contribute to this, but just wanted to throw it out there. BTW, does anyone know if Lucene can index existing database records (weblog entries)? I would imagine so. Would we want to index bookmarks and newsfeeds as well? The nice think about the Google search on this site is that it indexes everything. Anyone know of a java-based google-like search package?

Aqua Text - How to Scale Down.

I sent an e-mail to the author (Rick Yaeger) of the aqua-text tutorial last night asking how to scale down my text and reduce the fuzziness (or as Lance called it, "the gel toothpaste look"). Here is Rick's response:

There is an easy way to scale the effects of the style in Photoshop 7. In fact, that's what the command is called "Scale Effects..." and it is found under the Layers menu under the Layer Style submenu. Select the layer that has the effects you wish to scale and use this command to adjust those effects on a percentage scale with preview.

Another method is to create your type at 72pt and then flatten the image and scale it down 25%.

So I tried this, and it definitely helped. Here is the old one, and the new one. I've noticed that I do lack one thing as a Blogger - and that is following up on my previous posts, so here goes...

PDF Searching. Thanks to Vince Mastrantoni (no blog) and Greg Klebus for their e-mails. These will help me find my solution I'm sure. Thanks Guys.

The obvious answer would be Adobe Acrobat. But, this is not open source and if you have many files to search through, would be cumbersome. I tried to do a batch search using it, but was unable to. It appears as though this may not be supported or an option.

Another alternative is the Lucene search engine available at:

http://jakarta.apache.org/lucene/docs/index.html

Now, this by itself doesn't search PDFs. However, you can add a plug-in to do this. See here:

http://www.i2a.com/websearch/

Other stuff that might help is here:

http://www.etymon.com/

Also, the Google internet search engine searches and indexes PDFs and allows them to be displayed in text format. I'm not sure how or if this might help but I wanted to point this out to you.

Lastly, there is a PDF filter that Adobe provides that can be plugged into IIS index server. Tried looking for it but couldn't find it. Maybe you'll have better luck.

Vince

---
Found that filter. Gave up too soon:

http://www.adobe.com/support/downloads/product.jsp?product=1&platform=Windows

Vince

---
Raible, (I seem to have read your explanation about the name :)

pdftotext (forgot the source, you need to google it yourself). The name says it all -- convert your PDF to text and do the search.

It's a default viewer under Midnight Commander on my linux box, so I can view PDF's as text easily.

Open source, and probably cross-platform.

You can also try Ghostscript, it probably has something similar (but it's big, and you'd need to go through Postscript files anyway).

I hope I could help.

Regards,
-Greg

Also, I mentioned a while back that I didn't use Mail for OS X, but instead I use Entourage. Truth is, I've started using Mail (over Entourage) for a very simple reason, it's faster. When I click on the icon, it opens faster. I'm easy that way.

To follow up on my iSync post, I can't wait for the actual release. When I import all my vCards from Outlook XP, it changes the "mobile" phone number to "main" in my Address Book. Because of this, my contacts' mobile numbers never make it to my phone - doh! This same import works fine in Entourage. Also, I can't seem to 1) sync with .Mac or 2) publish my iCal to my webdav-enabled Tomcat server. Not that I really need to, but I like to experiment and see if it's possible. Manually changing my contacts' numbers from "main" to "mobile" in Address Book fixes the problem, but what a pain. And my phone seems to end up with a bunch of contacts I never put in there - iSync definitely has room for improvement. I'm looking forward to this update.

Last, but certainly not least, to follow up on where is Russ - it's great to have you back Russ, you were missed. You update your blog more frequently than any others I read - thanks.

In other news, I can see Phoenix being my default browser very soon. It's Mozilla without the extra stuff (i.e. Mail) that I don't use.

The Future of Roller.

To answer Dave's questions:

1) What do you want Roller to be?

I want it to be my personal portal, my weblogger, and my corporate website. I'm also interested in adding a wiki to my website for new projects. But I don't expect or need wiki functionality in roller. I'm planning on adding a new roller page with an iframe to a wiki such as Very Quick Wiki. The iframe gives me the ability to view another site/app with the current skin being used.

2) What are you interested in working on?

• Converting the editor UI to use Tiles instead of including header/footer JSP pages.
• Making the editor UI XHTML-compliant, while maintaining some backwards compatibility with Netscape 4.x.
• Adding the ability to collapse/expand folders for bookmarks.
• Adding search capability to weblog entries and entire site pages, either using Lucene or just 'like' SQL statements.
• Investigating XDoclet and Castor and how it's being used so I can use them on my next project.
• Adding an option for drop-down menus for editor UI (alternative to tabbed menu). I know there is a "Struts Menu" that has been converted to use CoolMenus, so maybe use this.

3) What features would you like to see added to Roller?

• The ability to search weblog entries.
• Ability to include an anchor name for each weblog entry, i.e. <a name="entryId..." id="entryId..."></a>. This would be convenient for other webloggers who are referencing your posts.
• Ability to use other blogging client-side software to post to roller, besides just w:blogger.
• Ability to edit (x)HTML via a web interface - in both Mozilla and IE.

I keep myself motivated by using Roller everyday as the engine behind this site. Thanks Dave, Roller has made the re-design of my website much easier, and it's opened my eyes to the world of weblogging. Good Stuff!