BigBlogZoo Forum: Last 35 Posts

zookeeper on "Incorrect tagged version of modules under HEAD in eclipse"

zookeeper — Sun, 19 Aug 2007 19:26:40 +0000

sorry for the late reply. I was on holiday.

Hmm. I am not sure. I will look into it later this week.

Does anyone else have an idea?

zookeeper on "Is BBZ the right app for this task?"

zookeeper — Sun, 19 Aug 2007 19:25:10 +0000

Sorry for the late reply. I was on holiday.

The app should work for you. However, Depending on you search terms, you may or may not get enough hits. In that case you may need to add some more sources.

What did you get stuck on? If you give me some more details I can try to help.

did you get the tutorial here : http://www.bigblogzoo.com/help6.html to work?

This is very similiar to what you want to do.

I am available now, so my reponse time should be faster.

But in response to your original question, I would say that bigblogzoo and in particular the mediaminer components is designed to do exactly what you want. There may be some bugs of course. And the reporting is lacking in some respects. However if you have some suggestions we are always listening and then of course if you want to implement some improvements that is also encouraged.

priyanshugupta21 on "Incorrect tagged version of modules under HEAD in eclipse"

priyanshugupta21 — Wed, 08 Aug 2007 07:28:27 +0000

I am using Eclipse version 3.0. Generally when we connect to a CVS repository through eclipse the "HEAD" (in the repositories explorer perspective) shows the latest tagged version of the module in repository and all the previous versions can be seen under the "Versions" heading in eclipse.
<p>
But my team started experiencing a problem for the past 1 week. Whenever we connect to CVS repository through eclipse it shows the oldest tagged version of the modules under the "HEAD" title instead of the latest version. And this is happening with all the modules in the repository. Please let me know the cause of this problem and what is the resolution for this.

prazim on "Is BBZ the right app for this task?"

prazim — Mon, 06 Aug 2007 02:24:17 +0000

Greetings!
I need to conduct essentially continual research for recent news articles that match specific key terms. I may also want to search for blogs that contain specific topics from time to time although I monitor several presently.

I have about 50 terms I want to search on, and desire to get results from worldwide media. They would be all separated by an "or", some are multi-word terms.

I tried using the program the other evening and got stuck almost immediately.

Thanks!
Sue

lgopinathan on "Eclipse 3.3 Migration"

lgopinathan — Sat, 28 Jul 2007 05:02:12 +0000

Thank you !!!!!!
It works like a charm now...
:-)

zookeeper on "Eclipse 3.3 Migration"

zookeeper — Fri, 27 Jul 2007 10:13:59 +0000

log4j:ERROR Ignoring configuration file [configuration\logging.conf].
java.io.FileNotFoundException: configuration\highlight.js (The system cannot find the file specified)

These errors are easy. The supplementary files are missing. You just have to download them from sourceforge and setup them up as details in the entry launching from source.

The only file which is crical is bigblogzoo.zoo (and it is only required in the bigblogzoo). The others are errors that should be ignorable, but in truth I have not gotten around to making the application robust enough to ignore them. Although this is a high priority.

The app will stil print to stdout the first time about some other missing files, but then it goes and creates them.

prozion on "Eclipse 3.3 Migration"

prozion — Fri, 27 Jul 2007 09:36:47 +0000

When I launch from Run Dialog, my launch file is:

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<launchConfiguration type="org.eclipse.pde.ui.RuntimeWorkbench">
<booleanAttribute key="append.args" value="true"/>
<booleanAttribute key="askclear" value="false"/>
<booleanAttribute key="automaticAdd" value="false"/>
<booleanAttribute key="automaticValidate" value="false"/>
<stringAttribute key="bootstrap" value=""/>
<booleanAttribute key="clearConfig" value="false"/>
<booleanAttribute key="clearws" value="true"/>
<booleanAttribute key="clearwslog" value="false"/>
<stringAttribute key="configLocation" value="${workspace_loc}/.metadata/.plugins/org.eclipse.pde.core/bigblogzoo.product (2)"/>
<booleanAttribute key="default" value="false"/>
<booleanAttribute key="includeOptional" value="true"/>
<stringAttribute key="location" value="${workspace_loc}/../runtime-bigblogzoo.product (2)"/>
<booleanAttribute key="org.eclipse.debug.core.appendEnvironmentVariables" value="true"/>
<stringAttribute key="org.eclipse.jdt.launching.PROGRAM_ARGUMENTS" value="-os ${target.os} -ws ${target.ws} -arch ${target.arch} -nl ${target.nl}"/>
<stringAttribute key="org.eclipse.jdt.launching.SOURCE_PATH_PROVIDER" value="org.eclipse.pde.ui.workbenchClasspathProvider"/>
<stringAttribute key="org.eclipse.jdt.launching.VM_ARGUMENTS" value="-Xmx512m -Xms256m"/>
<stringAttribute key="pde.version" value="3.3"/>
<stringAttribute key="product" value="bigblogzoo.product"/>
<stringAttribute key="productFile" value="\bigblogzoo-client\bigblogzoo.product"/>
<stringAttribute key="selected_target_plugins" value="com.ibm.icu,javax.servlet,javax.servlet.jsp,org.apache.ant,org.apache.commons.el,org.apache.commons.logging,org.apache.jasper,org.apache.lucene,org.apache.lucene.analysis,org.apache.xerces,org.apache.xml.resolver,org.eclipse.ant.core,org.eclipse.core.commands,org.eclipse.core.contenttype,org.eclipse.core.databinding,org.eclipse.core.expressions,org.eclipse.core.filebuffers,org.eclipse.core.filesystem,org.eclipse.core.filesystem.win32.x86,org.eclipse.core.jobs,org.eclipse.core.net,org.eclipse.core.resources,org.eclipse.core.resources.compatibility,org.eclipse.core.resources.win32,org.eclipse.core.runtime,org.eclipse.core.runtime.compatibility,org.eclipse.core.runtime.compatibility.auth,org.eclipse.core.runtime.compatibility.registry,org.eclipse.core.variables,org.eclipse.debug.core,org.eclipse.equinox.app,org.eclipse.equinox.common,org.eclipse.equinox.http.jetty,org.eclipse.equinox.http.registry,org.eclipse.equinox.http.servlet,org.eclipse.equinox.jsp.jasper,org.eclipse.equinox.jsp.jasper.registry,org.eclipse.equinox.preferences,org.eclipse.equinox.registry,org.eclipse.help,org.eclipse.help.appserver,org.eclipse.help.base,org.eclipse.help.ui,org.eclipse.help.webapp,org.eclipse.jdt,org.eclipse.jdt.compiler.apt,org.eclipse.jdt.compiler.tool,org.eclipse.jdt.core,org.eclipse.jdt.debug,org.eclipse.jdt.launching,org.eclipse.jface,org.eclipse.jface.databinding,org.eclipse.jface.text,org.eclipse.osgi,org.eclipse.osgi.services,org.eclipse.pde,org.eclipse.pde.build,org.eclipse.pde.core,org.eclipse.platform,org.eclipse.swt,org.eclipse.swt.win32.win32.x86,org.eclipse.team.core,org.eclipse.text,org.eclipse.tomcat,org.eclipse.ui,org.eclipse.ui.cheatsheets,org.eclipse.ui.forms,org.eclipse.ui.ide,org.eclipse.ui.ide.application,org.eclipse.ui.intro,org.eclipse.ui.intro.universal,org.eclipse.ui.navigator,org.eclipse.ui.navigator.resources,org.eclipse.ui.views,org.eclipse.ui.views.properties.tabbed,org.eclipse.ui.win32,org.eclipse.ui.workbench,org.eclipse.ui.workbench.texteditor,org.eclipse.update.configurator,org.eclipse.update.core,org.eclipse.update.core.win32,org.eclipse.update.ui,org.mortbay.jetty"/>
<stringAttribute key="selected_workspace_plugins" value="bigblogzoo,bigblogzoohelp"/>
<booleanAttribute key="tracing" value="false"/>
<booleanAttribute key="useDefaultConfig" value="true"/>
<booleanAttribute key="useDefaultConfigArea" value="true"/>
<booleanAttribute key="useProduct" value="true"/>
<booleanAttribute key="usefeatures" value="false"/>
</launchConfiguration>

Kent, you can compare it in file diff and see that there minimal differences, all related to directory names and similar things. Target plug-ins are the same and in the same amount :)

Maybe my problem is not in plug-ins? Honestly I had to comment out some piece of code in application: framework.MyActionBarAdvisor.fillMenuBar()
...
/*for (int i = 0; i < typesOfSearch.length; i++)
{
searchAlgolMenu.add(typesOfSearch[i]);
} */
...

because it depends on <eclipse-ide-installed-dir>/configuration/searchxml file and prints a null-pointer error when is not commented. I've created this file but it's empty.

May it cause the problems? Is the content of files: searchxml, highlight.js, logging.conf, bigblogzoo.zoo, searchcanals.scv, searchview.hts critical for running the application?

lgopinathan on "Eclipse 3.3 Migration"

lgopinathan — Fri, 27 Jul 2007 08:10:48 +0000

hey, I have all the plug ins required but now I have these errors
1.
log4j:ERROR Ignoring configuration file [configuration\logging.conf].
java.io.FileNotFoundException: configuration\highlight.js (The system cannot find the file specified)

2.
Exception in thread "Thread-2" java.lang.NullPointerException

3.
log4j:ERROR Could not read configuration file [configuration\logging.conf].
java.io.FileNotFoundException: configuration\logging.conf (The system cannot find the file specified)

zookeeper on "Eclipse 3.3 Migration"

zookeeper — Thu, 26 Jul 2007 20:39:17 +0000

ok, hmm maybe for some reason validate is not working for you guys.

In the Run Dialog, go down the list list of plugins and manually tick org.eclipse.core.runtime(3.3.100.v20070530).

And try to launch, there is a good chance some other plugins are still missing, so keep adding until it works (add the next plugin that it says is missing). You can periodically press add required, as soon enough you will hit the missing one and add required will begin to work.

Below is my launch configuration. The very long line is the list of required plugins.

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<launchConfiguration type="org.eclipse.pde.ui.RuntimeWorkbench">
<booleanAttribute key="append.args" value="true"/>
<booleanAttribute key="askclear" value="false"/>
<booleanAttribute key="automaticAdd" value="false"/>
<booleanAttribute key="automaticValidate" value="false"/>
<stringAttribute key="bootstrap" value=""/>
<booleanAttribute key="clearConfig" value="false"/>
<booleanAttribute key="clearws" value="true"/>
<booleanAttribute key="clearwslog" value="false"/>
<stringAttribute key="configLocation" value="${workspace_loc}/.metadata/.plugins/org.eclipse.pde.core/bigblogzoo.product"/>
<booleanAttribute key="default" value="false"/>
<booleanAttribute key="includeOptional" value="true"/>
<stringAttribute key="location" value="${workspace_loc}/../runtime-bigblogzoo.product"/>
<booleanAttribute key="org.eclipse.debug.core.appendEnvironmentVariables" value="true"/>
<stringAttribute key="org.eclipse.jdt.launching.PROGRAM_ARGUMENTS" value="-os ${target.os} -ws ${target.ws} -arch ${target.arch} -nl ${target.nl}"/>
<stringAttribute key="org.eclipse.jdt.launching.SOURCE_PATH_PROVIDER" value="org.eclipse.pde.ui.workbenchClasspathProvider"/>
<stringAttribute key="org.eclipse.jdt.launching.VM_ARGUMENTS" value="-Xmx512m -Xms256m"/>
<stringAttribute key="pde.version" value="3.3"/>
<stringAttribute key="product" value="bigblogzoo.product"/>
<stringAttribute key="productFile" value="\bigblogzoo\bigblogzoo.product"/>
<stringAttribute key="selected_target_plugins" value="com.ibm.icu,javax.servlet,javax.servlet.jsp,org.apache.ant,org.apache.commons.el,org.apache.commons.logging,org.apache.jasper,org.apache.lucene,org.apache.lucene.analysis,org.apache.xerces,org.apache.xml.resolver,org.eclipse.ant.core,org.eclipse.core.commands,org.eclipse.core.contenttype,org.eclipse.core.databinding,org.eclipse.core.expressions,org.eclipse.core.filebuffers,org.eclipse.core.filesystem,org.eclipse.core.filesystem.win32.x86,org.eclipse.core.jobs,org.eclipse.core.net,org.eclipse.core.resources,org.eclipse.core.resources.compatibility,org.eclipse.core.resources.win32,org.eclipse.core.runtime,org.eclipse.core.runtime.compatibility,org.eclipse.core.runtime.compatibility.auth,org.eclipse.core.runtime.compatibility.registry,org.eclipse.core.variables,org.eclipse.debug.core,org.eclipse.equinox.app,org.eclipse.equinox.common,org.eclipse.equinox.http.jetty,org.eclipse.equinox.http.registry,org.eclipse.equinox.http.servlet,org.eclipse.equinox.jsp.jasper,org.eclipse.equinox.jsp.jasper.registry,org.eclipse.equinox.preferences,org.eclipse.equinox.registry,org.eclipse.help,org.eclipse.help.appserver,org.eclipse.help.base,org.eclipse.help.ui,org.eclipse.help.webapp,org.eclipse.jdt,org.eclipse.jdt.compiler.apt,org.eclipse.jdt.compiler.tool,org.eclipse.jdt.core,org.eclipse.jdt.debug,org.eclipse.jdt.launching,org.eclipse.jface,org.eclipse.jface.databinding,org.eclipse.jface.text,org.eclipse.osgi,org.eclipse.osgi.services,org.eclipse.pde,org.eclipse.pde.build,org.eclipse.pde.core,org.eclipse.platform,org.eclipse.swt,org.eclipse.swt.win32.win32.x86,org.eclipse.team.core,org.eclipse.text,org.eclipse.tomcat,org.eclipse.ui,org.eclipse.ui.cheatsheets,org.eclipse.ui.forms,org.eclipse.ui.ide,org.eclipse.ui.ide.application,org.eclipse.ui.intro,org.eclipse.ui.intro.universal,org.eclipse.ui.navigator,org.eclipse.ui.navigator.resources,org.eclipse.ui.views,org.eclipse.ui.views.properties.tabbed,org.eclipse.ui.win32,org.eclipse.ui.workbench,org.eclipse.ui.workbench.texteditor,org.eclipse.update.configurator,org.eclipse.update.core,org.eclipse.update.core.win32,org.eclipse.update.ui,org.mortbay.jetty"/>
<stringAttribute key="selected_workspace_plugins" value="bigblogzoohelp,bigblogzoo"/>
<booleanAttribute key="tracing" value="false"/>
<booleanAttribute key="useDefaultConfig" value="true"/>
<booleanAttribute key="useDefaultConfigArea" value="true"/>
<booleanAttribute key="useProduct" value="true"/>
<booleanAttribute key="usefeatures" value="false"/>
</launchConfiguration>

by the way I can't find anything out right now about how to import a luanch configuration. but please tell me if you have tried this and if it works.

prozion on "Eclipse 3.3 Migration"

prozion — Thu, 26 Jul 2007 18:20:38 +0000

I have the same error:
Bundle initial@reference:file:plugins/org.eclipse.core.runtime_3.3.100.v20070530.jar/ was not resolved. It's after I do "Validate Plug-ins" and launch application from bigblogzoo.product.

But if I try to launch it from Run Dialog, then I receive no errors, see the splash picture, and then loading(init) process stops. If to look on the javaw process, that I guess is responsible for running the application, then it doesn't take CPU time. Just dead. Is it something with Windows?

Also I can try to launch application in debugger. The point where program freezes is following: WorkbenchAdvisor.run(). It calls display.readAndDispatch() which is never set initDone[0] to true (it's initialization as I guess?) so it never breaks a while(true) loop.

zookeeper on "Eclipse 3.3 Migration"

zookeeper — Wed, 25 Jul 2007 20:01:09 +0000

this is really odd.

I just downloaded Eclipse for RCP/Plug-in Developers - Windows (153 MB) for here :
http://www.eclipse.org/downloads/

I followed all of the steps and it worked.

Remember that the first time it should fail then you need to add the required plugins.

What happens when you press validate?

Something else you can try is to go to the main plugin -> right mouse click -> pde tools -> and try update class or organize manifests.

I have a sneaky feeling that you have the wrong version of eclipse (not the rcp one) or after it fails the first time you are not correctly adding the required plugins.

Let me explain what is happening, when you press "Launch an Eclipse Application" the very first time, it tries to create a launch configuration and launch, but that launch configuration is not valid, so then you have to go and make it valid by adding the required plugins, then once you do that you when you try to launch again and it should work.

Sorry it is not working. Not worry we will figure it out. Be sure to tell me what is says when you press validate.

lgopinathan on "Eclipse 3.3 Migration"

lgopinathan — Wed, 25 Jul 2007 13:05:43 +0000

Hey,

I am out of luck again..............I tried these steps again for the bigblogzoo........Haven`t had any luck as yet........
I am still getting this error..
Bundle initial@reference:file:plugins/org.eclipse.core.runtime_3.3.100.v20070530.jar/ [3] was not resolved.
What can I do to get rid of that ?

clueless
lg

zookeeper on "Feature Under Development : Category Synchronization"

zookeeper — Mon, 23 Jul 2007 21:56:15 +0000

Sure I would be happy to give some more info.

Basically I am going to presume you have the BigBlogZoo downloaded and installed.

The tree of on the left side is our collection of rss newsfeeds and blogs. It is classified using the DMOZ classification system. You can find more info about dmoz at dmoz.org. Dmoz has probably a million or so websites classified, we currently only have 60,000 feeds, so as you can see we use a subset of dmoz. However sooner or later someone is going to want to add a feed and the exact category will be missing. So what do we do? Well we need to somehow "deprune" a directory. Pruning is what in cvs you call removing empty directories, so I call it depruning meaning that you add back empty directories.

Now we can "deprune" in one of two ways a) like I have outlined above b) or a database driven solution where we have the complete dmoz tree on a database and when a user wants to deprune they select the directory and the client sends a message to the database conaining the present directory and the database sends the collection of subdirectories to the client.

If you have any questions, feel free to ask.

ravimeetsu on "Feature Under Development : Category Synchronization"

ravimeetsu — Mon, 23 Jul 2007 21:03:31 +0000

Hi there,

I would like to know more about the project before I put my hands up to lead it.

Could you please point me to where to start reading to get more information about the project. I do not have much understanding about the project from above.

Any information is appreciated. Thanks.

Ravi

zookeeper on "Database"

zookeeper — Fri, 20 Jul 2007 20:56:31 +0000

These all good points.

You are absolutely right about the security issue it is something that needs to be sorted out. And you are also right, the client can read & write, I will change that post, oops ;). I also see that we need to synchronize the database. I have nothing against a servlet solution, but I would like to avoid unneccessary complexity if possible. Can the client not lock tables in order to synchronise?

I don't see any false conclusions.

I have a question for you. You can see that in ConnectionHandler the login credentials are currently hard coded. I added these credentials "by hand" to the database. How does this happen programmitically?

I have added below some same data from the tables so that you can get a feel for the data, Users and Votes tables are currently empty.

--
-- Daten für Tabelle Changes
--

INSERT INTO Changes VALUES (74186, 1, '2005-08-23 00:00:00', 0, NULL);
INSERT INTO Changes VALUES (74185, 1, '2005-08-13 11:33:41', 0, NULL);
INSERT INTO Changes VALUES (74184, 1, '2005-08-12 21:05:49', 0, NULL);
INSERT INTO Changes VALUES (74183, 1, '2005-08-12 13:40:17', 0, NULL);
INSERT INTO Changes VALUES (74182, 1, '2005-08-12 13:39:22', 0, NULL);

--
-- Daten für Tabelle Feeds
--

INSERT INTO Feeds VALUES (60, 334, 'A Girl Apart', '', 'http://agirlapart.typepad.com/index/index.rdf', '', 0, 1, 0);
INSERT INTO Feeds VALUES (61, 339, 'RSS-Newsfeeds bei IzyNews - Kultur & Events.Ya', 'IzyNews - Nachrichten, Weblogs und RSS in der Mail. Tagesaktuell einsortiert, schnell zu lesen und 100% spam-frei. Das grÃ¶Ã?te deutsche RSS-Verzeichnis auf http://izynews.de/dir/. Nur bei IzyNews: E', 'http://izynews.de/de/dir/rss/Kultur%20%26%20Events', 'Ger', 0, 1, 0);
INSERT INTO Feeds VALUES (62, 339, 'RSS-Newsfeeds bei IzyNews - TV & Medien.YahooG', 'IzyNews - Nachrichten, Weblogs und RSS in der Mail. Tagesaktuell einsortiert, schnell zu lesen und 100% spam-frei. Das grÃ¶Ã?te deutsche RSS-Verzeichnis auf http://izynews.de/dir/. Nur be', 'http://izynews.de/de/dir/rss/TV%20%26%20Medien.Yah', 'Ger', 0, 1, 0);
INSERT INTO Feeds VALUES (63, 340, 'We Read Shite (So You Don''t Have To)', 'A blog that lists updates to fanfic recs pages', 'http://www.smallvillefanfic.com/te.rss', 'Eng', 0, 1, 0);
INSERT INTO Feeds VALUES (64, 342, 'Elsa''s Live Journal', 'Elsa''s Live Journal - LiveJournal.com', 'http://www.livejournal.com/users/elsaf/data/rss', 'Unk', 0, 1, 0);
INSERT INTO Feeds VALUES (65, 342, 'Furl - Michael Giles''s Archive', 'Furl archive.', 'http://www.furl.net/members/mgiles/rss.xml', 'Unk', 0, 1, 0);

--
-- Tabellenstruktur für Tabelle Nodes
--

INSERT INTO Nodes VALUES (1809, 'Abstract', 1808);
INSERT INTO Nodes VALUES (1810, 'Roleplaying', 1795);
INSERT INTO Nodes VALUES (1811, 'Science_Fiction_and_Fantasy', 1810);
INSERT INTO Nodes VALUES (1812, 'Message_Boards', 1811);
INSERT INTO Nodes VALUES (1813, 'Science_Fiction', 1812);
INSERT INTO Nodes VALUES (1814, 'Play-By-E-Mail', 1811);
INSERT INTO Nodes VALUES (1815, 'Science_Fiction', 1814);
INSERT INTO Nodes VALUES (1816, 'VGA_Planets', 1815);
INSERT INTO Nodes VALUES (1817, 'Television_and_Movies', 1810);
INSERT INTO Nodes VALUES (1818, 'Star_Trek', 1817);

prozion on "Database"

prozion — Fri, 20 Jul 2007 14:12:10 +0000

Oops, I've just downloaded some records from syndi4_dynamicbigtree database. (Using Runtime->Databases in NetBeans). Before I thought in 'Feeds' we have records that correspond to some kind of articles, where 'Feeds.Description' field contains the article's text. Now I see I'm wrong: 'Feeds' is the table of links to rss/... channels. And articles (bundles of articles) by itself stay at their generic location pointed by 'Feeds.Url'. Learn the basics... :)

Anyway, let me tell you what I think about database handling. Kent, you wrote, that client can not read the database. But if to consider client as code, running on the client side, then I see the client reads and writes to the database directly. For example the calls sequence for adding a feed into the database: *:AddFeedAction.(...).run() -> JobHandler.scheduleUpdateJob() -> DatabaseLiason.addFeed() -> *:FeedDetails.addFeed() -> <u>*:java.sql.PreparedStatement.executeQuery()</u>

where * denotes some instance(object) of the class.

Hence, users of BigBlogZoo client theoretically could send random queries to the database (by changing the code of client, or using some sql client) and that is the issue of security. Also it raises the question about synchronization of users requests to the database. While all part of the code is on the client side, it would be hard to control when and what requests will come to the database, and in what order to process these multiusers requests. I don't realize clearly for a while, where namely could be a problem in synchronization, but I feel it easily could happen in future.

What if to move some part of the code to the server? For a given example, the code are the classes DatabaseLiason and FeedDetails (classes from the package com.generatescape.database). Then scheduleUpdateJob() will send operation parameters to a servlet, and the servlet will address that data further to DatabaseLiason.addFeed() method. Also it's interesting to try to implement communication using RMI to call addFeed() directly, as a variant.

By the way, then for each connection it would be possible to organize a session, so client logins just once and then sends its requests without preliminary authorizing on each call.

That's my proposal :)
I don't develop this idea further because I admit I could be wrong in some conclusions about client working. I'd like to hear your remarks.

zookeeper on "Eclipse 3.3 Migration"

zookeeper — Thu, 19 Jul 2007 14:08:46 +0000

hi-ya!

Have you tried these insructions for SearchSaver?
I did and I didn't have any problem. Do an update and try these instructions, and if that doesn't work tell me.

There may still be an using with the BigBlogZoo and SearchSaver being open in the workspace. If you do have both of them open, select the BigBlogZoo plugins, press the right mouse and close them and then try all of these steps again. If that is the case, and this works, tell me so that I can fix it. Oh and lastly make sure that your eclipse 3.3 is the RCP / plugin developer version.

Until Later,

kent

lgopinathan on "Eclipse 3.3 Migration"

lgopinathan — Thu, 19 Jul 2007 07:00:58 +0000

Hai !

I have a whole load of warnings and errors when I launch bigblogzoo and the search saver.
error warnings sound something like this
Bundle initial@reference:file:plugins/org.eclipse.update.configurator_3.2.100.v20070615.jar/ was not resolved.

Could you tell me how I can resolve that?

zookeeper on "Feature Under Development : Feed Annotations and Ratings"

zookeeper — Mon, 16 Jul 2007 19:41:06 +0000

alrighty, sounds like a plan. I should have some time a bit later this week and I will try and give some detailed ideas and/or suggestions. good luck.

jova59 on "Feature Under Development : Feed Annotations and Ratings"

jova59 — Sun, 15 Jul 2007 19:15:59 +0000

Ok for your remarks.
I will start tomorrow with the basic(and easy) action :
- Add a rate (1 to 5 stars) => tinyint
- Add a note (limited to 255 characters for the moment).

I haven't think about showing all the rates but only the pondered rate (in the table feeds).
If we don't need the pondered rate, then we don't need a transaction.
I waiting for your(or others) suggestion.

I will look better with the problems of utf-8 and the asynchronous load of the tree (rates, annotations) later and I told you what I have found to resolve this.

I will follow your suggestion of using a local database for my testings.

zookeeper on "Feature Under Development : Feed Annotations and Ratings"

zookeeper — Sun, 15 Jul 2007 17:30:39 +0000

The tables look good.

NOTE VARCHAR(255) -> 255 is probably too small. How much a hit do we suffer if we do not define the lenght?

I think for the rating we could use TINYINT but the user is only presented with say 5 choices (or ten), for example 1 to 5 stars.

One thing when you do create the tables them we need to make sure that annotations can support utf-8.

Also I would suggest, but you can do it as you like, to maybe set up derby locally for a test environment.

sounds good. a) One user can add only one rate for a given feed.
sounds good. b) A user can modify his rate for a feed.
sounds good. c) A user can add more than one note to a given feed.
sounds good. d) A user can modify his annotations for a feed.

sounds good - Add a rate => a new action "Rate this feed" in the contextual menu
=> the action will only be enabled if one and only one feed was selected.

sounds good - Add a note => a new action "Add a note to this feed" in the contextual menu => the action will only be enabled if one and only one feed was selected.

But how do we show all the notes for a given feed ? I was thinking about showing votes and annotations as nodes in the trees. We may have performance issues, but the idea is that when you open a node we send asynchronous message to the database and try and get any votes or annotations for that feed. What do you think?

Do we need actually need the rate in the feeds table? I am not sure anymore.

Hmm sounds unneccessary complicated, using a cron job. If we need it I suppose we could use a transaction. SQL is not my strong point, does anyone else have an idea?

PS you are welcome!

jova59 on "Feature Under Development : Feed Annotations and Ratings"

jova59 — Sat, 14 Jul 2007 16:31:45 +0000

I was too busy to work on this feature this last weeks so now it's ok,I have time to think at this feature.
So here is some propositions and some questions.

1) Propositions of tables :
Ratings :
CREATE TABLE RATES (
FKEY INT(11) NOT NULL, -- feed key
UKEY INT(11) NOT NULL, -- user key
RATE TINYINT(3) NOT NULL, -- rate : 0 to 100 ?
PRIMARY KEY (FKEY,UKEY));

Annotations :
CREATE TABLE ANNOTATIONS (
AKEY INT(11) NOT NULL auto_increment, -- annotation key
FKEY INT(11) NOT NULL, -- feed key
UKEY INT(11) NOT NULL, -- user key
NOTE VARCHAR(255), -- note for this feed
PRIMARY KEY (AKEY));

2) Questions :
Are you ok with this propositions :
a) One user can add only one rate for a given feed.
b) A user can modify his rate for a feed.
c) A user can add more than one note to a given feed.
d) A user can modify his annotations for a feed.

zookeeper on "Feature Under Development : Category Synchronization"

zookeeper — Sun, 08 Jul 2007 11:50:50 +0000

The BigBlogZoo uses the DMOZ classification schema, however only a subset of it. What we need though is someway of synchronizing the BigBlogZoo tree with the complete schema.

To do this completely on the client side, the client side needs the complete schema to be available. That is no problem, but it needs to be stored in an efficient way. The compressed version here: http://rdf.dmoz.org/rdf/ file:structure.rdf.u8.gz is already 66 megs. But 90% of that is xml fluff.

So if we can save the contents of this file fluff free and if we access the data in a reasonably quick way we are there.

People already moan about the size of the download so I would really like to set a goal 6 megs on the file size. The data is however in utf-8 so we have to be careful how we save it, we need to preserve the encoding.

The idea is quit simple :

we convert and possibly compress this heirarchical data into some format
we provide accessors to this data. The accessor is dead simple, I give you a parent directory, you have to give me the children.

A naive complete implementation could be as follows:

You read structure.rdf.u8.gz and create a hashmap.
Your key is the parent and your value is children.
Then you serialize the map. Then you optionally zip the serialized map. When you want to read you dezip, deserialize, pass in your key and get your value(s).

By the way parsing the rdf might not be the best idea, sometimes dmoz xml is not even well formed. You can try, but sometimes I find I just have to it by reading in the text and capturing strings.

zookeeper on "Feature Under Development : Lucene Search"

zookeeper — Sun, 08 Jul 2007 11:43:34 +0000

Currently we use lucene to match searches. However we do not serialize a lucene search base. In other words after an ad hoc search the search base is not kept, but thrown away. It would be nice to store that search base.

zookeeper on "Eclipse 3.3 Migration"

zookeeper — Sun, 08 Jul 2007 11:07:40 +0000

We have now migrated to Eclipse 3.3.

Eclipse 3.3 offers a rich set of new feautures that we can easily integrate into the BigBlogZoo. As with any migration this does pose some problems. Currently we are still having a problem with the launch configuration, but for the most part I believe this this can be solved by doing the following:

get a cvs update and recompile
From the product plugin - overview tab - press Launch an Eclipse Application (it will probably fail)
then in eclipse go to the run menu -> open run dialog-> plug ins tab-> add required plug-ins -> then validate.
From the product plugin - overview tab - press Launch an Eclipse Application (this time it should not fail

After validation the you should see something like this:

If you have any problems give me some details I will try to sort them out.

zookeeper on "Eclipse Communication Framework ECF"

zookeeper — Fri, 06 Jul 2007 17:27:07 +0000

There is something brand new in 3.3 that I think would make a great addition, it is called the Eclipse Communication Framework. It would have some effects on the user validation scheme we have created.

What does it do? Well at the most basic level it enables, chat, but then it should also enable shared browsing, and eventually sharing of workspace components. The whole big idea of the BigBlogZoo is a collaborative network, so we need this baby.

If we set up an ECF server then there is probably no real effect on our user concept. But if we need to use google or yahoo as communication providers, then the user will have to login twice or least provide two sets of credentials.

Here is conversation I add with the ECF People.
####################

I am running an open source project called the bigblogzoo. In our next release we are concentrating on the community aspect of the application, and we would like to integrate ecf, therefore it would be great if you could
give us some pointers.

Ideally we would like to do the following:
a) have a chat client
b) be able to share a View ie a TreeViewer. This sharing could be somehow messaging based, ie an encoded message is parsed on the reciever's side
which fires a message to open the tree to a certain node.
c) be able to cobrowse. I can't seem to get this to work or find documentation for this.
d) do all of this with out hosting an ecf server, ie be able to use providers like google or yahoo

#### ANSWER ####

I think the best way for you to go, given 'd' is to use the ECF datashare API (for 'a', 'b' and 'c') for the data you wish to exchange
(i.e. for b and c).

'a' is a little problematic, however, given 'd' (if you mean multiuser chat, rather than 1-1 chat). Google talk does not (yet) support XMPP
multiuser chat (although I've heard rumors that they intend to), and the ECF yahoo provider does not (yet) implement the ECF chat APIs (we have
an enhancement request for that here
https://bugs.eclipse.org/bugs/show_bug.cgi?id=146955 please add yourself to the cc to 'vote' for this bug)

But in any event, you should be able to use the datashare API to exchange messages for the TreeViewer and to do URL co-browsing.

The basic pattern for getting a datashare channel (a messaging channel for arbitrary data), is:

// Get a container instance...somehow
IContainer container = <get through some means...see below>

// Get the IChannelContainerAdapter
IChannelContainerAdapter channelContainerAdapter =
(IChannelContainerAdapter)
container.getAdapter(IChannelContainerAdapter.class);

// Create a named channel ("mychannel") to use for your messaging
IChannel channel =
channelContainerAdapter.createChannel(IDFactory.getDefault().createStringID("mychannel"),listener,
null);

// Use the channel.sendMessage() to send messages to channel, and
//listener.handleChannelEvent to receive from to channel. Remember that
// the thread that calls listener.handleChannelEvent(event) is unlikely
// to be the UI thread, and it can be called at any time.

Note that the datashare API is currently implemented by the XMPP provider, the skype provider, the ECF generic provider, and the JMS
provider. We haven't yet had the chance to implement it with the Yahoo provider.

Here are the Datashare API docs, with pointers to test code, javadocs, and source code: http://wiki.eclipse.org/ECF_API_Docs#Datashare_API

Now, you may be wondering...how do you get a hold of an IContainer instance (to start the above process)? There are essentially three
ways...and which is desireable to you depends upon what kind/level of UI integration you want with the existing UI (some, none, full):

1) Query the IContainerManager. There is an OSGI service for ECF called IContainerManager. It provides access to all the previously-created
IContainer instances. You access it in the normal way for OSGI services (e.g.)

ServiceTracker tracker = new
ServiceTracker(bundleContext,IContainerManager.class.getName(),null);
tracker.open();
IContainerManager manager = (IContainerManager) tracker.getService();
IContainer [] containers = manager.getAllContainers();
// look for container in containers[] that you are interested in

2) Create a new IContainer instance and connect it yourself. You would want to do this if you want to have users connect to an account that only shows your own user interface (rather than integrates with the existing roster view and/or chat UI). To create/connect an IContainer,
use the ECF ContainerFactory:

IContainer container =
ContainerFactory.getDefault().createContainer("<type>");

where <type> is "ecf.xmpp.smack" (XMPP/Google Talk) or "ecf.yahoo.jymsg"
for yahoo.

3) Get an IContainer instance from the existing UI. For example, if the user connects to their google talk account, using the existing ECF UI,
this will bring up the "Contacts" View. This view is implemented in org.eclipse.ecf.presence.ui/MultiRosterView class. With Eclipse 3.3 UI, it's possible to add UI to existing views (e.g. menu items, etc). From
these additions, it's possible to access the selected roster elements (i.e. IRoster and IRosterEntrys), and get the IContainer from (e.g.)
IRosterEntry (your buddy). So, for example, if the current selection is your buddy:

IRosterEntry rosterEntry = (IRosterEntry) getSelection();

IContainer container =
rosterEntry.getRoster().getPresenceContainerAdapter().getAdapter(IContainer.class);

This allows full integration with the MultiRosterView/Contacts view...if
this is desired.

There are two examples of doing this in the
org.eclipse.ecf.presence.collab.ui plugin:

http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.ecf/plugins/org.eclipse.ecf.presence.collab.ui/?root=Technology_Project

This uses the org.eclipse.ui.menus.menuContribution extension point to add dynamic menu items to the
org.eclipse.ecf.presence.ui.MultiRosterView context menus. Also, see the URLShare and ViewShare classes in this plugin as examples of adding simple protocols for remote URL launching and remote View opening.

Finally, I think that Marcelo Mayworm (an ECF committer) was going to do some work on modifying the Eclipse built-in browser so that it would
allow automatic co-browsing rather than the user-initiated co-browsing that is there now. You may want to work together on this. And Kent if
you can contribute back any of these applications to ECF, this would be most welcome.

Thanks,

Scott

zookeeper on "What is the BigBlogZoo?"

zookeeper — Fri, 06 Jul 2007 17:19:00 +0000

The first diagram is correct, however the client feed repository is technically not a database. The client feed directory is a swing modelled tree (package com.generatescape.newtreemodel ). The model is swing but the gui components are swt.

The flow between client and server database directional? The client calls the central repository and asks for updates (see class SyncAction) and also sends new submissions.

Regarding the second diagram, let me explain how reaggregation works (see class ArchiveAction ). Also note that this feature is available to anyone, however I do not really document it as there is probably a fair bit of hardcoding that makes it work for me.

The spider runs and it scans for new articles.
It then creates dated feeds for each category. So for example if I ran the spider today and I had not run it for a couple of days, the spider would create ARCHIVE_2007_07_05.xml, ARCHIVE_2007_07_04.xml, ARCHIVE_2007_07_03.xml, etc for the category Top/Business/Accounting/.
Then I have a seperate program that transforms these xmls in html using xslt and creates a zip. This program is not included currently in the client and is not on the server, so if we were to include it in the system diagram it would be outside of the system.
Then I upload this onto the server and unzip it.

The bigblogzoo.com server you could call the aggregation server, but note that it really is a passive file server.

Currently the server really only takes care of deltas. If you add a feed it gets sent to the server and then other clients can synchronize on the database, in other words they can then download new feeds.

What is planned is that the server can:

register moves from one category to the other
store user's ratings of feeds
store user's annotations of feeds
store user data

There is a forum entry that has the current database schema.

By design the client is the real worker.

By design there are two servers. There is one server which hosts simple stuff like html, php and javascript pages (bigblogzoo.com). Then there is another server that hosts the database and supports servlets (syndicatescape.com).

I was and still am considering moving everything to one server. I have decided if the syndicatescape.com host is not providing backups or if backups are expensive then we will probably do everything on the bigblogzoo.com server. The bigblogzoo is a virtual private server server using suse.

Ideally I would like two servers, because I think the solution is more scalable on a low budget, and more fault tolerant.

One thing that I want to know, is it possible to use jdbc to add a user to a database?

prozion on "What is the BigBlogZoo?"

prozion — Thu, 05 Jul 2007 19:55:49 +0000

The second chart with more details comes here:

I split the server module on the feeds aggregation server and the database server. Also I added browser (thin client) that allows to read aggregated feeds from the server (bigblogzoo) using google maps and timeline.

Red elements denote components and data flows that I suspect exist somewhere but I failed to detect corresponded source code (that I checked out from the project CVS repository)

prozion on "What is the BigBlogZoo?"

prozion — Thu, 05 Jul 2007 19:24:21 +0000

Do I understand the system correctly? Here is the first (and simplest) chart:

The cubes stand for the system modules, that are server and client module. Each module contains a database(repository) with feeds. Arrows show the data flow direction. Bi-directional arrow between two databases means that local database can read the data from server database as well as write its own data back to the server database.

Yellow rectangles with a bent corner are remarks. Number shows the possible amount of modules. 1 means server exist just in single copy. 1..* means we assume the number of clients is from 1 to indefinite value

But maybe I'm wrong even with this simple conception. While learning the code from CVS I decided to assign the feeds aggregation function on the diagram to the client module, because I couldn't find the code responsible for feed aggregation on the server side. (Probably I've just missed it)

Blue-violet rectangle denotes server that publishes feeds (in RSS, Atom or other XML feeds format but not HTML)

zookeeper on "Database"

zookeeper — Mon, 02 Jul 2007 19:02:09 +0000

Here is our current database. I will add some more comments soon.

// FKey = Feed Key
// UKey = User Key
// NewPath = A vote for a newpath

CREATE TABLE Votes (
UKey int(11) NOT NULL default '0',
FKey int(11) NOT NULL default '0',
NewPath char(255) NOT NULL default '',
PRIMARY KEY (UKey,FKey),
KEY FKey (FKey),
KEY NewPath (NewPath),
KEY UKey (UKey)
) ENGINE=MyISAM DEFAULT CHARSET=latin1;

CREATE TABLE Users (
UKey int(11) NOT NULL auto_increment,
UserId char(30) NOT NULL default '',
PRIMARY KEY (UserId),
KEY UKey (UKey)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=1 ;

CREATE TABLE Nodes (
NKey int(11) NOT NULL auto_increment,
Name char(255) NOT NULL default '',
Parent int(11) NOT NULL default '0',
PRIMARY KEY (NKey),
KEY Parent (Parent)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=58166 ;

CREATE TABLE Feeds (
FKey int(11) NOT NULL auto_increment,
ParentNode int(11) NOT NULL default '0',
Name char(255) NOT NULL default '',
Description char(255) default NULL,
Url char(255) NOT NULL default '',
Language char(30) NOT NULL default '',
Rate tinyint(3) unsigned default NULL,
Hits smallint(6) default NULL,
Changes smallint(6) default NULL,
PRIMARY KEY (FKey)
) ENGINE=MyISAM DEFAULT CHARSET=latin1 AUTO_INCREMENT=74762 ;

// This table tracks changes whether they are moves or adds
// currently only action 1 (add) is implemented

CREATE TABLE Changes (
FKey int(11) NOT NULL default '0',
Action tinyint(4) NOT NULL default '0',
ChangeDate datetime NOT NULL default '0000-00-00 00:00:00',
FormerParentNodeId int(11) default NULL,
Heritage char(255) default NULL,
PRIMARY KEY (FKey)
) ENGINE=MyISAM DEFAULT CHARSET=lati

zookeeper on "Why do we use SWT and not Swing?"

zookeeper — Thu, 28 Jun 2007 20:07:29 +0000

There are many articles and debates about which one is "better". I am not going to detail the plus points of each, but just answer why we use it. The answer is not so much about SWT per se, but more about SWT + JFace + RCP as a whole.

Sure SWT has native widgets which perform better and give the user the familiar look and feel, but that is not the whole story. The next layer of abstraction is JFace. JFace takes these widgets and combines them into resuable patterns, for example TreeViewers, TableViewers, Wizards, etc. Sure "sometimes" swing has things like this but not always, does swing have virtual tables? Not that I know of. Does swing have a standard set of widgets for displaying error messages? Does swing have a way of presenting preferences to the user? Then comes the next layer of absraction, the Rich Client Platform. Client side java has always failed to be able to deliver a complete client experience. That is what the RCP is, a way of delivering an executable product, that is extensible and complete. It offers searchable help, splash screen, pluggable lazy loaded actions, branding, etc.

zookeeper on "Feature Under Development (SearchSaver): RAP Version"

zookeeper — Mon, 25 Jun 2007 16:53:49 +0000

Eclipse has a new ajax harness called RAP. Many of the SWT widgets can be used in a web browser. If we could port SearchSaver into a RAP Application we would have a very powerful meta search engine.

zookeeper on "Feature Under Development (SearchSaver): Better Selection of Search Engines"

zookeeper — Mon, 25 Jun 2007 16:51:22 +0000

The current method for selecting search enegines is not very user friendly. We need to implement a new way of selecting which search engines to use.

zookeeper on "Feature Under Development : PHP Template Work"

zookeeper — Sat, 23 Jun 2007 20:51:56 +0000

We need to unify the html elements of the BigBlogZoo Website into one look and feel.

The project is about:

writing a php template for our bbPress forum
writing a php template for our wordpress mu web site
later customizing those templates more so that we incorporate the BigBlogZoo application into the web more closely

Currently we have a php / javascript / html / web based BigBlogZoo and a desktop application BigBlogZoo.

The Web Based BigBlogZoo has three looks and feels:

the homepage http://www.bigblogzoo.com (Wordpress Blog)
the forum which you now reading (BBPress Forum)
the categorized feeds - for example http://www.bigblogzoo.com/Top/Business/Associations/Associations.html

We need to take the graphical elements and styles of the categorized feed pages (see item 3) and incorporate them into BBPress and Wordpress php templates.

So in other words the idea is to make this page and http://www.bigblogzoo.com/ look like
http://www.bigblogzoo.com/Top/Business/Associations/Associations.html , and doing this by creating BBPress and Wordpress PHP templates.

BBPress and Wordpress are made by the same people so the template syntax is probably very similiar. The developer is also free to incorporate any features from Wordpress, BBPress or the BigBlogZoo that they wish to incorporate into the templates. For example wordpress widgets, or the results of querys on the BigBlogZoo database, etc.

You can find bbpress themes documenation here:
http://bbpress.org/documentation/themes/

This how to create a WordPress theme:
http://codex.wordpress.org/Theme_Development

The first step would be to download and install bbpress on a server. Then to experiment with the themes. It might even be a good idea to modify an existing them rather than creating one from scratch.

The first priority would be the template for the forum.

zookeeper on "What is the BigBlogZoo?"

zookeeper — Wed, 20 Jun 2007 18:17:59 +0000

The BigBlogZoo has two faces, the web version and the desktop application.

The online BigBlogZoo is a semantic web. By no means is it "the semantic web". However, it is an experiment in building an interconnected web that can be traversed time wise and location wise.

How does it work? Articles from the BigBlogZoo application are regularly aggregated, uploaded to our server and a timeline or google map is generated.

What about the BigBlogZoo Application? It is a sort of semantic news aggregator. 60,000 rss/atom feeds have been categorized using the DMOZ classification schema. In many respects it is like a news aggregator with 60,000 feeds preloaded and classified. Basically it is a RSS reader with a couple of twists, grouping similar content into over 60,000 "channels" that users can search, browse and syndicate.

The next release is planned to extent the "community" aspect of the BigBlogZoo in order to get more people to submit feeds and to help classify them.

For the next big release we plan to create better and more crowdsourcing.

Wikipedia defines crowd sourcing as "Each member of the crowd puts forth an insignificant contribution to the total outcome, but the grand total of these contributions amounts to a considerable difference."

One of the main ideas of the BigBlogZoo is to create world´s largest directory of news feeds with help from our user base. Currently users can submit their newsfeed to a collection of DMOZ categorized feeds. We would like to further develop these idea by a system of voting, whereby users can also categorize feeds.

There is also data mining component called the MediaMiner. With MediaMiner you can do categorized crawling and reaggregation of the BigBlogZoo.

Results results are saved as HTML, .pdf or XML (Atom or RSS). The software is designed for people who need breaking news. Typically search engines require at least a week before they have the latest database update. With MediaMiner you get the results instantly.

Also useful for research, MediaMiner can scour the web to find out what customers are saying about a particular company, and because the results can be saved as exportable .pdf format, it is useful for generating reports and sharing the data.