Gossamer Forum
Home : Products : Gossamer Links : Discussions :

Extended Admin Reviews

Quote Reply
Extended Admin Reviews
I have a client who will eventually need a links system that does this ...

... or some of this ...

* Supports extended reviews by admin or admin-appointed editors.
These would not be the normal brief descriptions or editor's
picks notes, but detailed evaluations of the value of a web
site to professionals -- up to 2000 words. All links would
have reviews, and they would have a common format with headings
and paragraphs.

* Online editing facility with login and passwords.

* Users would be able to add comments, with optional moderation.

* Editors would be able to add and edit private notes for each
link and category.

* Each link could be assigned to multiple categories at the
editor's discretion.

* Multiple logical hierarchies (eg: links could be assigned
to one or more categories in a Regional hierarchy, plus one
or more categories in a topical hierarchy).

* Regular link validation, with error codes stored in the
database.

* Dependent links (ie: where a very large web site can be split
into several internal sites, each with their own review).


Is it likely that we would ever be able to use Links SQL for
this sort of scheme? We hope so ...



Quote Reply
Re: Extended Admin Reviews In reply to
all useful but some already there

extended reviews... easy.. just use the detailed page.

online editing.. well admins and mods can do that as well as people who add their links can edit theirs

add comments.. i think alex is going to release a review plugin soon but unfortunatly suggested it would not be moderated, something i think that needs to be changed

editor private notes. Not sure how hard that would be to add, probably not very just an extra column in the table somewhere but have to ask someone else on that

links is multiple categories. Easily done in category browser

link validation. Pretty sure that is there already.

dependant links. How do you want to do this? need to explain more but i am sure it could be done.

Conclusion, you can pretty much do everything already thus is the power of links :)

http://www.ASciFi.com/ - The Science Fiction Portal
Quote Reply
Re: Extended Admin Reviews In reply to
It sounds like you are trying to add the functionality of DMOZ (although I haven't looked in awhile. I was an original editor there for 2 main categories, but quit over a year ago when their "human edited" directory became nothing more than a human-spidered directory without any real 'editing' allowed).

Some of the functionality of what you need is already in there. The rest (except one or two subtlties) are things I have thought about adding in already.

The problem: These are all new mods, and new code. If you had patience, in the next 6 months to a year, many of these things would probably be built in, or available as a string of mods by a number of people.

If you need it sooner, contact me privately. Ballpark: $2000-$3000 to get 80% of that working in the next 20-30 days. $8-10,000 to get it all working within the next 30-60 days (debugging aside).

>> * Dependent links (ie: where a very large web site can be split
>> into several internal sites, each with their own review).

This is something I need to seriously discuss with Alex, once the new code is released. We have a string of messages on this issue going back almost 2 years. I don't work for GT, so I have not made any mods that would circumvent their licensing (at least publically -- Alex knows I've been playing with certain ideas on my own server). The "prefix" option on the database seems to be Alex's first step in that direction. One solution might be to add a server/site licensing option that would allow running multiple databases off one installation, that are related by a core database, sort of like a star-cluster formation, with each of the main sub-cats a separate list of categories/links -- and maybe even user authorizations.

Anyway, depending on how you intend to want this part set up, it may require additional Links licensing, or it might need to wait until the library routines are updated to handle multiple connections of this sort. Right now, the core logic is still "one-dimentional" in how it looks at the links database. To do what I think you are asking, the logic has to be upgraded to handle a 3-dimentional relationship (for a visual picture, think of 3-d spreadsheets).






PUGDOGŪ
PUGDOGŪ Enterprises, Inc.
FAQ: http://pugdog.com/FAQ


Quote Reply
Re: Extended Admin Reviews In reply to
pugdog not sure bit i think you misunderstood the last bit.. sorry if you havn't. I think he wants to be able to do:

www.domain.com/part1 and then www.domain.com/part2 as differnt listings/reviews.

http://www.ASciFi.com/ - The Science Fiction Portal
Quote Reply
Re: Extended Admin Reviews In reply to
or not even that. It is something i have thought about when you want www.domain.com listed in one category and the same www.domain.com in another category but with different reviews and no problem of duplicate check. Perhaps that is what is being talked about as opposed to licensing issues?

Anyway... any new features to make it more like dmoz.com would be good news, that it definitely the example software to follow really and the only directory program so far that is still beter than links.

http://www.ASciFi.com/ - The Science Fiction Portal
Quote Reply
Re: Extended Admin Reviews In reply to
Many, perhaps all, of the functions that we will need can
be hacked into existing Links flatfile. I've done 75% of it
using modifications from Resources.

However, the point I wanted to make is that there is
possibly a substantial niche market for professional
gateway systems that the Links authors could tap into.
But they would need to think about the real needs of such
a market. It's easy to justify investing a few hundred $$$
on good quality software, but not having to spend a few
thousand more on modifications -- it should work straight
out of the box.

The nearest Open Source system I have found so far is
ROADS. Check these URLs -- they explain what I need far
better than I can myself ...

http://www.ilrt.bris.ac.uk/roads/
http://www.ukoln.ac.uk/metadata/roads/what/
http://www.ilrt.bris.ac.uk/roads/who/

But ROADS actually does much more than I need. It's also a
nightmare to install (I have tried), unsuited to virtual
servers, and does not have the extensive user community
that keeps Links going so well. Although it addresses my
needs, it is not what I am looking for.

The WWW Virtual Library at http://vlib.org/ is a
closer model for me than the Open Directory Project -- or
perhaps to be more accurate, something that combines the
best features of Links, ODP and the Virtual Library, but on
a relatively small scale. For the project I have in mind,
500 core links would be a realistic maximum -- small for
Links as we know it, but with reviews the database would
be at least 2.5 Mb.


I didn't explain "dependent links" very well. Here is an
example ...

Suppose I have a gateway for cancer treatments ... there
are numerous web resources that should be reviewed, some
good, some mediocre, many that are trivial, and a few that
are misleading ... a good case here for professional
evaluations by editors who have knowledge of the subject
matter and of how the Internet works.

One of the best resources is CancerNet at
http://cancernet.nci.nih.gov/.
Like many of its kind, it is a very large site with
hundreds of subtopics. A lot of those would merit
separate reviews, so my idea would be to have a primary
review, and embedded within it would be the links to the
subtopic reviews, which in turn would have links back
to the main review.

The issue is not duplication, but quality control.
If CancerNet has been reviewed as a source of reliable,
up to date information, then the dependent sub-sites
would 'inherit' some aspects of the main review, with
new editor input at the lower level.

Quote Reply
Re: Extended Admin Reviews In reply to
The 75% that can be hacked into Links 2.0 are probably already in Links SQL. It's the other 25% that is the problem.

Very often 90% of the mod can be made very easily, but the last 10% takes 90% or more of the effort.

The more features you add into a program, the more complex it becomes to install, maintain, administer, and even develop. "out of the box" means something different to each person.

For most sites, the program works "out of the box" except for customizing the templates, and that is something each site really _needs_ to do.

Details, extra information, extra categories and indexes, extra levels, extra interactions, all of that is site-specific. It's not easily built into a product.

I need to run an image site. My links are images. I have to make changes to each release to make it work the way I need it to. To have it run this way "out of the box" would mean it wouldn't work for all the people who _are_ running "links" sites with jumps to external links.

Also, the pricing on Links SQL is very, very reasonable considering the amount of work that went into it, and how much abstraction they put on top of the other free-ware/open-source products -- apache, perl, mysql, etc -- to make it easier to use.

>> a market. It's easy to justify investing a few hundred $$$
>> on good quality software, but not having to spend a few
>> thousand more on modifications -- it should work straight
>> out of the box.

I try to restrain myself from getting too hyper, but people often ask for $1000's in mods to be posted to the forums, as if it's the normal course of activity. It's not.

If you want to customize the product, learn perl, sql, apache, Unix, Tcp/IP, etc, and do it yourself. If you want to design your own product, you are free to do so. You can also use open source products such as ROADS or MINIVEND to do similar things. You can also spend significantly more on other products, and still have to modify them -- or deal with excess features that bog down performance.

Links SQL 2.0 is probably a 1/4 to 1/2 million dollar investment should a company try to re-develop it themselves. It would also take a minimum of 4 or 5 people working 6 months to do it. $500 is a pretty reasonable price for that basis of software.

Given that, the cost of $10,000 in modifications is _NOT_ unreasonable either.

You are getting the development engine -- such as Dbase or Oracle -- for a reasonable price. It's now up to you to develop the application.

How much do applications cost? Targeted niche applications start at $500 for "hobby" type applications, and go up to $25,000 or $50,000 for applications that do things like run jewelry stores or handle process flow. Full-blown applications can run millions of dollars -- and they may be built on the "inexpensive" foundation of a standard database and "free" Linux, Perl and Apache.

"open source" and "free software" doesn't mean it's not going to cost you anything. All it means is that the code is open, and you are "free" to use it and alter it to your needs. Many companies got listed on the NASDQ supporting and servicing "free" software.

I'll shut off the flame thrower here, but I've been programming commercially on and off since 1982. I have contributed back to the software community in many ways over the years, and operate the FAQ and support the program here as a give back to all those who helped me along all those years. I don't expect people's time to be "free" for the asking.

And as for what GT is doing, I'd personally rather have a stable engine, and set of building blocks that _can_ be built upon, than a set of fluffy features that don't work as expected, and which become a night mare to maintain as a site is deployed.

The longer a program is out, the more that will be developed to work with it, but development, unless paid for, is only at the speed the "hobbiests" do it.

As you've indicated, you aren't willing to fund the costs of development of new features, and neither are most people. To that end, it will take time to get a library of extra software, modifications, or plug ins.

These things don't just "happen", someone has to do them. That person has to invest time and resources. "Just for fun" doesn't cover the house payments or the food bills.



PUGDOGŪ
PUGDOGŪ Enterprises, Inc.
FAQ: http://pugdog.com/FAQ


Quote Reply
Re: Extended Admin Reviews In reply to
OK pugdog ... actually, I agree with both your arguments and your ballpark pricing, which seems very reasonable to me. I suppose I should have said so ...

Explaining it to the client would not be so easy. But that would be my job.

At this stage I am evaluating all possible options. Links SQL is still my favored solution, and one of the main reasons is the large user community and available expertise. My role in this project is advisory. The client does not have to take my advice -- and I would not be surprised if they allocate a budget which is less than what is needed to do the job properly. I have to be prepared for a full range of outcomes.

The client here is a non-profit organization. The issue is not whether they are prepared to fund the cost of new features, but whether they can do it.