My family tree needs more wood and less sap.
 |
pnTNG and Google Indexing |
|
|---|---|---|
| steve30000 | Posted: 09.12.2005, 12:27 | |
|
registered: Nov. 2005
Posts: 12
Status: offline last visit: 01.10.06 |
It appears to me that search engines are not going to handle the indexing of TNG within Postnuke very well since it is contained in an IFRAME. Is this a founded concern? How do others handle this situation? My initial thought is that the IFRAME code should contain an alternate link to TNG for browsers (or search engines) that don't handle IFRAMES. Thoughts? Steve |
|
|
|
|
| Wendel | Posted: 10.12.2005, 10:36 | |
|
registered: Jun. 2005
Posts: 82
Status: offline last visit: 03.05.08 |
Steve, Yes, I came to a similar conclusion a couple of weeks ago. I think there are two problems: 1. Use of the IFRAME itself – it is not clear to me if the spiders traverse the IFRAMES 2. Inside the IFRAME the URLs used are direct URLs to TNG – by passing PostNuke. So if it indexed, the direct path to the TNG will be indexed, and you will not see any PostNuke header/footer/sideblocks, etc. So far I have looked at 3 approaches: 1. The “TNG Nuke Add-on v1.8.1” embeds TNG without the IFRAMES. It works by using a hook left in from the early days of PostNuke. But to use it, I believe pnTNG would loose much of its ease in setup --- in effect pnTNG would be a duplicate of this work. I am also concerned that since this is an old way of accessing PostNuke modules, the writers of PostNuke are more likely to remove support for it as PostNuke progresses over time. If that happens, then another solution would be needed anyway. 2. I tried to wrap the TNG output in a different way. In php (the language pnTNG is written) there is an “include” statement. There is also a mechanism to capture the output in a buffer. So I wrote up a version that did that. The problem came in a quirk with php in that “global variables” are not carried into the include. It turns out TNG uses “global variables” extensively throughout the code. I tried a work around by identifying all “global variables” that are in TNG and had it working about 80% --- but the list became so long and as soon as TNG changes even one variable then pnTNG would have to change as well. 3. I began experimenting with still using the IFRAME but changing all the URLs to go through PostNuke. Currently, when you get into pnTNG, it creates an IFRAME and then everything with TNG happens inside the IFRAME --- that is why you do not see the PostNuke blocks refresh while in TNG. I am trying to make it to where for each URL clicked in pnTNG, instead of staying inside the IFRAME, it goes back to the site and generates a new page with PostNuke and a new IFRAME. There might be other ways too. If you know of a way, I am open to input. I also need to do some more reading on IFRAMEs and will scavange around the PostNuke forums. I can’t believe this is the only program trying to do this. |
|
|
|
|
|
| Wendel | Posted: 10.12.2005, 14:31 | |
|
registered: Jun. 2005
Posts: 82
Status: offline last visit: 03.05.08 |
OK, I found a postnuke module that does what we need (and more.) I even hacked it to work with TNG. So I have an example to go from now. But it also has a lot of stuff that pnTNG doesn't need. So I need to tear it apart and figure out what is important. The module is NukeWrapper found at http://users.tpg.com.au/staer/Downloads.html It looks like a pretty handy thing to have. You can wrap text, other sites, etc. edited by: Wendel, Dec 10, 2005 - 01:34 PM |
|
|
|
|
|
| steve30000 | Posted: 11.12.2005, 18:24 | |
|
registered: Nov. 2005
Posts: 12
Status: offline last visit: 01.10.06 |
I actually already use a wrapper on my site for a block that I have linked to the USA Today crossword puzzle. The wrapper I use is PostWrap that is available at www.spideon.com. After looking at it though, I see that it uses IFRAMES as well. It looks as though NukeWrapper is the way to go. Is there a reason that you want to rework pnTNG and not just propose the use of NukeWrapper for the main TNG access and use pnTNG for the block add ons that you have (Updates, Birthdays, etc.)? My original thought on the IFRAME use was to just put a link to the TNG site in between the IFRAME tags. From my research, it appears as though if the search engine cannot follow the IFRAME it will read the non-frame compatible browser part. So if there is a link to TNG in that portion of the code, then the search engine will be able to follow it. But as you said, the results in the search engine will no longer contain the postnuke header, etc. I will probably try to install the NukeWrapper in my site and see how it works. |
|
|
|
|
|
| Wendel | Posted: 11.12.2005, 18:48 | |
|
registered: Jun. 2005
Posts: 82
Status: offline last visit: 03.05.08 |
Steve, The reason to embed the NukeWrapper into pnTNG is to allow for the sharing of usernames between PostNuke and TNG. Before handing over to TNG, the new wrapper needs to get the TNG user data and set up some session variables so TNG knows who is logged in. It also creates the TNG users if needed (and if allowed.) Along the way, I think it will also be easier to set some of the NukeWrapper settings automatically, so users are not having to hand edit the file. Right now though, for the most part, I'm taking the NukeWrapper file and inserting my changes in one area (for easier upkeep and updates later on.) Then of course, I have the side blocks to update as well. If you want, I can send a work-in-progress file to show you what I have so far. I still do not have the administration pages working (and there are a few other quirks.) - Wendel |
|
|
|
|
|
| steve30000 | Posted: 11.12.2005, 19:16 | |
|
registered: Nov. 2005
Posts: 12
Status: offline last visit: 01.10.06 |
Good points. I'm always logged in so I didn't think about those issues. If you would like someone to test things for you, feel free to send me files. Steve |
|
|
|
|
|
| steve30000 | Posted: 05.02.2006, 13:59 | |
|
registered: Nov. 2005
Posts: 12
Status: offline last visit: 01.10.06 |
I've done some looking around and it appears that any site that is run with Postnuke is not getting the genealogy section indexed at all in google Besides, your site and my site, I checked ccmerritt.com and gilles-family.de (the other sites that are listed on Darren's site as postnuke sites using TNG). The sites have extensive indexing but do not include any genealogy information. Here's what I'm wondering. Since the module is using login information from postnuke and is set to require login, is this causing google to not be able to access the genealogy? I tried spidering my site with a simulator (http://www.searchengineworld.com/cgi-bin/sim_spider.cgi) and I get valid URLs during the search but the spider results are showing individuals as Unknown so therefore the spider is seeing every page as identical and not indexing. What do you think? Steve |
|
|
|
|
|
| Wendel | Posted: 19.02.2006, 13:30 | |
|
registered: Jun. 2005
Posts: 82
Status: offline last visit: 03.05.08 |
Steve, I just found this site last night: www.snakelab.com They state: PostNuke web sites produce dynamically generated pages that are invisible to most search engines. When the search engine sends out Crawlers/Spiders to index your page it will not follow any URL that contains a parameter like (?) or (%). If they see a link like this on your home page, http://www.snakelab.com/index.php?module=ContentExpress&func=display&bid=19&btitle=Company&mid=1&ceid=37 they will disregard everything after the question mark (?). To the spider, the link looks like something it has already indexed. http://www.snakelab.com/index.php If you take a look at SnakeSEO, just be careful and read the forums first. One of the things it does has caused Google to ban folks. He was using a javascript to redirect pages -- a no-no by Google. I'm not sure if you had found this or not. I'm going to start reading up on the mod_rewrite that can be placed in an .htaccess file. It seems others in the Postnuke world are doing this too. They call it using "Short URLs". It seems it is an option in the Xanthia configuration page. From my little bit of reading, it looks like we could come up with a nice way of having Google friendly URLs that get transformed into what PostNuke and pnTNG need. Also, great work on the site mapper! edited by: Wendel, Feb 19, 2006 - 01:29 PM |
|
|
|
|
|
| Guest | Posted: 19.02.2006, 14:43 | |
| Unregistered User | I had read about SnakeSEO somewhere before in the postnuke forums. Everyone over there thought it was wonderful but I never really explored since my main interest was getting the genealogy portion of my website spidered... hence the sitemap creator tool. I have submitted a sitemap that goes just to my TNG installation and bypasses postnuke. On my TNG pages then, I have the search and home pages redirecting to my main postnuke site. I figured if I can get them to a page using google, then I can get them back to my genealogy site wrapped in postnuke if they do a search or go to the home page. Not the most elegant way but it was the only way I could think of after realizing the troubles of getting anything inside postnuke spidered by google. Although if you look at what google had indexed on my site previously (site:hooverfamily.com), it had no problem going through the postcalendar module and finding all the information so I don't know what to think of that. I don't think anyone understands how google gets its information! :) | |
|
|
|
|
| steve30000 | Posted: 19.02.2006, 14:46 | |
|
registered: Nov. 2005
Posts: 12
Status: offline last visit: 01.10.06 |
I have a problem of forgetting to login before I post here. Anyway, that last message was from me if you couldn't tell. Steve |
|
|
|
|
|
Users online:
This list bases on the users active in the last 10 minutes
Powered by pnForum Version 2.6