Author Topic: I can code for you guys  (Read 432 times)

djtrickdog

  • Larvae
  • *
  • Posts: 23
Re: I can code for you guys
« Reply #20 on: July 11, 2009, 05:54:59 PM »
hey i fixed  the corners so they are smoother, ill upload in a few hours! i am still hoping for a list of request from you vesp, where are you bud?

Vesp

  • Administrator
  • Foundress Queen
  • *****
  • Posts: 3,130
Re: I can code for you guys
« Reply #21 on: July 12, 2009, 07:47:40 AM »
Don't call me bud.
I'm just going to keep it how it is for now, and I'll give you suggestions on the updated version.
Bitcoin address: 1FVrHdXJBr6Z9uhtiQKy4g7c7yHtGKjyLy

djtrickdog

  • Larvae
  • *
  • Posts: 23
Re: I can code for you guys
« Reply #22 on: August 02, 2009, 03:37:50 AM »
heres the updated version guys:
http://127.0.0.1/djtrickdog01/Sleek/

Vesp says he likes his site how it is though so this is...done i guess haha.

heisenberg

  • Dominant Queen
  • ****
  • Posts: 268
Re: I can code for you guys
« Reply #23 on: August 02, 2009, 04:52:10 AM »
heres the updated version guys:
http://127.0.0.1/djtrickdog01/Sleek/

Vesp says he likes his site how it is though so this is...done i guess haha.

I think that looks pretty sexy.  I'd vote for it if there were a poll.
I spent all my money on booze and hookers, the rest I wasted - Charles Bukowski

Vesp

  • Administrator
  • Foundress Queen
  • *****
  • Posts: 3,130
Re: I can code for you guys
« Reply #24 on: August 02, 2009, 09:24:28 PM »
That does look really good, but I like the simplicity of the homepage.
Also when the idea of these designs were made, it was an open forum and was going to have an archive, library, etc.

Now I'd rather have the "archive" - That is synthesis, extractions, how to grow certain things, etc be in the publications section of the forum.

I now think keeping this site small with only the front page and FTP accounts visible to the public/google spiders is a good thing. Allowing only members to be able to view the forums contents is probably a good thing. Similar to how WD and other such forums have done.

However, if other people want the forum changed into something different, please post or PM me about it. I'd like to do what is best for the forum/members.




Bitcoin address: 1FVrHdXJBr6Z9uhtiQKy4g7c7yHtGKjyLy

wellbie

  • Larvae
  • *
  • Posts: 25
Re: I can code for you guys
« Reply #25 on: August 06, 2009, 02:28:49 AM »
u can ride in a lair jet to get where ya going .sure !

then u can get in old faithful ,some thing u know that works ,dependable   that will get u to the same destination, may bee not as fast but she get u there for sure !

some times slow and easy is the best way to go !

and from what i have seen so far u all on the right path and some of the selections will pay off in the future ,hope the team work's out for ya!
« Last Edit: August 06, 2009, 02:32:47 AM by wellbie »

shroomedalice

  • Guest
Re: I can code for you guys
« Reply #26 on: November 25, 2009, 01:36:36 PM »
personaly I think the owner of the site should do all the code unless you have known the person for a long time or
you have checkers of the server side code.

with php perl java and jscript its real easy to do some realy naughty things and the nature of the site just screams
for some one to get there claws in it for reasons other than advancing chemistry.

I can code very well but I dont think I would code anothers site to be honest.

maby if you find a coder it would be best to have three checkers to make sure its all the way it
should be.

then again having to many people know the insides of the server side code is also very dangerous.

php is not very hard to learn and sql insert selects and unions are not hard either.

there is an old saying if its not broke dont fix it.

Vesp

  • Administrator
  • Foundress Queen
  • *****
  • Posts: 3,130
Re: I can code for you guys
« Reply #27 on: November 25, 2009, 06:32:58 PM »
Quote
there is an old saying if its not broke dont fix it.

Yeah nice saying, I probably live by it. :P
Bitcoin address: 1FVrHdXJBr6Z9uhtiQKy4g7c7yHtGKjyLy

no1uno

  • Global Moderator
  • Foundress Queen
  • *****
  • Posts: 681
Re: I can code for you guys
« Reply #28 on: March 14, 2010, 10:45:16 AM »
Anyone who can code "very well" is invited to help with a library  ;D
"...     "A little learning is a dang'rous thing;
    Drink deep, or taste not the Pierian spring:
    There shallow draughts intoxicate the brain,
    And drinking largely sobers us again.
..."

badger

  • Larvae
  • *
  • Posts: 24
Re: I can code for you guys
« Reply #29 on: March 14, 2010, 02:28:51 PM »
I'm happy to help but don't have the time to do the whole thing myself. PHP is what pays my rent.

My advice as far as a library goes is to start with a mockup of the user interface for the library. Photoshop is fine, HTML+CSS is better (though I or someone else can turn a Photoshop design into proper HTML+CSS). Get comments and suggestions on that, and tweak it until you have an interface that it seems like people will be happy with, then code the backend to use the frontend.

no1uno

  • Global Moderator
  • Foundress Queen
  • *****
  • Posts: 681
Re: I can code for you guys
« Reply #30 on: March 14, 2010, 08:36:15 PM »
PHP? Cool, I'm happy enough with the XHTML/CSS/Jscript - you get the PHP working with the backend (ie. MySQL) and that will take some serious time off my hands ;D

We need (IMHO) a decent AJAX type application, so the innerHTML responses to fill the <div id=content> is going to be replaced on a regular basis, in fact everytime the response comes in ;)

Have a look at XAJAX... Seems perfect for this aplication

In terms of the aesthetics of it, a single picture (logo), the rest is just a completely .css/XHTML top/bottom bar, side bar and search... Light as fuck and loads like a dream. Match that with AJAX and there is no need for a reload, just narrow down the searches by (1) using suggest by making an array of ALL authors; and/or using multiple dropdown menus to restrict searches into the predefined format, Subject, Topic & Sub-topic, to give a reasonable amount of responses (all visible with fuck all scrolling) on that same page with no reload.

« Last Edit: March 14, 2010, 08:48:34 PM by no1uno »
"...     "A little learning is a dang'rous thing;
    Drink deep, or taste not the Pierian spring:
    There shallow draughts intoxicate the brain,
    And drinking largely sobers us again.
..."

badger

  • Larvae
  • *
  • Posts: 24
Re: I can code for you guys
« Reply #31 on: March 14, 2010, 09:03:12 PM »
We need (IMHO) a decent AJAX type application, so the innerHTML responses to fill the <div id=content> is going to be replaced on a regular basis, in fact everytime the response comes in ;)

Yes, though ugh, no .innerHTML, that way lies madness. jQuery and DOM manipulation all the way.

Quote
In terms of the aesthetics of it, a single picture (logo), the rest is just a completely .css/XHTML top/bottom bar, side bar and search... Light as fuck and loads like a dream. Match that with AJAX and there is no need for a reload, just narrow down the searches by (1) using suggest by making an array of ALL authors; and/or using multiple dropdown menus to restrict searches into the predefined format, Subject, Topic & Sub-topic, to give a reasonable amount of responses (all visible with fuck all scrolling) on that same page with no reload.

Yep, I'm familiar with the technique (it's the "Live Search" pattern on ajaxpatterns.org, http://ajaxpatterns.org/Live_Search). Done it before.

So, the user interface that you have in your head? Mock it up in Photoshop (or the Gimp or whatever) and we can turn that into HTML/CSS, then we'll work out the schema and queries for the database and nail it all together.

no1uno

  • Global Moderator
  • Foundress Queen
  • *****
  • Posts: 681
Re: I can code for you guys
« Reply #32 on: March 19, 2010, 10:56:40 AM »
So, let me get this straight... All you want from me is a basic XHTML page, with the relevant CSS and you are willing to help with the design of the DB/PHP?

Cool ;)
"...     "A little learning is a dang'rous thing;
    Drink deep, or taste not the Pierian spring:
    There shallow draughts intoxicate the brain,
    And drinking largely sobers us again.
..."

badger

  • Larvae
  • *
  • Posts: 24
Re: I can code for you guys
« Reply #33 on: March 19, 2010, 11:19:07 AM »
Yes. I'm already fiddling around with Drupal and Solr, which is Apache's full-text search engine for both text and rich documents (e.g., Word or PDF). Don't know if it handles DjVu, but that's probably solvable by converting the file.

no1uno

  • Global Moderator
  • Foundress Queen
  • *****
  • Posts: 681
Re: I can code for you guys
« Reply #34 on: March 19, 2010, 01:25:09 PM »
I don't see why we need to bother with drupal, etc.

I have just modified a basic, generic 3 column layout, the html and css files are attached.

With some simple jscript it would be possible to use onclick with the top links and slightly modify the .css file to have 5 tabs - Home, Subject, Author, etc. with the relevant XHTML built into the existing code (to handle the onclick). If the user is an admin/mod, then they would have access to a couple of locked tabs, namely upload and modify...

The bottom links can be either disabled where needed (and just use one for copyright/legal and another for contact)...

If we did that, I'd suggest just putting a series of dropdown menus on each tab (left column) and instructions (right column) and the results being all in one div in the center column

Of course, the Home tab, would have some basic settings in the dropdowns (left column, the amount dependant upon access/privileges) and a login box in the top of the right column.

That minimizes the page load dramatically and will cut down on the php/mysql/ajax significantly.

As to using DOM modification, it hardly seems worthwhile when there is only one div that needs modifying once the dropdown menus are utilized.

I can't see why we need to use heavy-duty search engines to go through files, the files won't even be in the DB, only the url/uri to them and their abstract need be in the db.

The database upload form would be in a tab, accessible only if a person has the requisite privileges, and the modify entry/etc. tab would be the same.

We'd simply modify the response(s) based on user privileges, anyone who isn't registered, plus anyone who isn't authorised, wouldn't get to see anything but the first three tabs, and the responses would only have links to full text versions of papers if they are hosted/already in the public domain. Basically, this would be the science search site to rule them all, we just lift the abstracts from JACS/Wileys/Elsevier/etc. and then link to them. No copyright issues whatsoever, they have that stuff on display already, we'd actually be helping them.

If people are authorised, then if the fulltext is available, their response(s) will have the link thereto as well as the link to the abstract and the abstract...

It would be something that is very useful for people studying and the potential earnings from google admiles is phenomenal.

The filesharing behind the scenes would be off-limits to all but the chosen few and could be arguably classed as 'fair use', given the restricted access thereto, within a known group of people, with common interests and goals, not distributed for commercial gain.

PS It is also possible to solve the entire tabs thing in pure .css/xhtml, which would simplify shit even further and allow people to use the site to the fullest with jscript off.
« Last Edit: March 19, 2010, 01:48:39 PM by no1uno »
"...     "A little learning is a dang'rous thing;
    Drink deep, or taste not the Pierian spring:
    There shallow draughts intoxicate the brain,
    And drinking largely sobers us again.
..."

badger

  • Larvae
  • *
  • Posts: 24
Re: I can code for you guys
« Reply #35 on: March 19, 2010, 01:41:30 PM »
I don't see why we need to bother with drupal, etc.

Oh, it's not necessary, but I do think a CMS will make certain things easier.

Quote
I have just modified a basic, generic 3 column layout, the html and css files are attached.

Thanks, I'll build some stuff off them later today.

Quote
With some simple jscript it would be possible to use onclick with the top links and slightly modify the .css file to have 5 tabs - Home, Subject, Author, etc. with the relevant XHTML built into the existing code (to handle the onclick). If the user is an admin/mod, then they would have access to a couple of locked tabs, namely upload and modify...

Yep, quite familiar with that sort of thing (both jQuery and scriptaculous have some nice tabbing code already written).

Note that once you get into different user roles (admin/mod/regular user/anonymous/whatever), letting a CMS do that work for you makes code easier to maintain.

Quote
As to using DOM modification, it hardly seems worthwhile when there is only one div that needs modifying once the dropdown menus are utilized.

For active search (ie, bringing up results in realtime based on what a user types in a search box) the code ends up much cleaner -- you have the backend return a list of JSON objects containing the information about the results, then some frontend javascript to turn those results into their own little divs, and CSS will style them automagically.

Quote
I can't see why we need to use heavy-duty search engines to go through files, the files won't even be in the DB, only the url/uri to them and their abstract need be in the db.

With Solr you don't put the files in the db. Solr runs as its own server; you index the files that you want full search enabled on, and you ask the Solr server for results based on its index.

Certainly we can start with just the information in the abstracts, but I think we'll get richer results if we add full-text searching. That can be done in a second pass, though.

no1uno

  • Global Moderator
  • Foundress Queen
  • *****
  • Posts: 681
Re: I can code for you guys
« Reply #36 on: March 19, 2010, 01:54:42 PM »
I see what you are saying, but I prefer simple, light, and fast...

There is no beating it - the amount of time and the sheer number of people who would be directed to this site (and I argue that it needs its own url) would be hell on server load

Plus, if we are only using abstracts that beats the hell out of searching the various publishers sites, we have everything people want, organised and with links to where to get it...

On the backend, the php for the various privileges would quite possibly require some serious thought

As to using .css to style the responses, this is a site I built a while ago, just playing around with using .css/xhtml (see frames are back in xhtml2.0 btw?)...

That would minimise the page space if users are restricted to searching by subject, topic, subtopic (in dropdowns), thereby avoiding the annoying as piss problems that come from having to scroll down huge pages to find shit. Simple .css onclick and the abstract etc. are displayed.
"...     "A little learning is a dang'rous thing;
    Drink deep, or taste not the Pierian spring:
    There shallow draughts intoxicate the brain,
    And drinking largely sobers us again.
..."

badger

  • Larvae
  • *
  • Posts: 24
Re: I can code for you guys
« Reply #37 on: March 19, 2010, 10:12:54 PM »
I see what you are saying, but I prefer simple, light, and fast...

There is no beating it - the amount of time and the sheer number of people who would be directed to this site (and I argue that it needs its own url) would be hell on server load

I really don't think it's going to be that bad. We're talking about pre-indexing a lot of data (if we go the Solr route) specifically so that users can search through a lot of rich content on a fine-grained level. I use PubMed a lot, and one thing I've noticed is that just searching for terms that appear in the abstracts (which is what PubMed's keyword search does) often doesn't help me as much as being able to search through the full text of the article would ... but we're not talking about anywhere near the amount of documents that PubMed has to index.

But anyway, my point is that an index of a PDF archive is just like a *database* index: it retrieves data very fast. We wouldn't be searching *publishers'* sites at all; we'd be indexing their content in order to point people to it.

This also gives us scalability, in the event that the archive becomes popular enough to need more than one machine -- the Solr server can be moved to another box, trivially.

I'm downloading the entire Rhodium archive right now, and will index it on my local installation, then stress-test it a bit.

Quote
On the backend, the php for the various privileges would quite possibly require some serious thought

Authentication is easy to get wrong. That's part of why I prefer a thoroughly tested system and would rather not reinvent that wheel.

Oh, and don't forget we'll also have to deal with stuff like pagination, another reason to use a CMS or templating framework.

Quote
As to using .css to style the responses, this is a site I built a while ago, just playing around with using .css/xhtml (see frames are back in xhtml2.0 btw?)...

I really like that layout -- that's very well done, and your code is cleaner than that of some professionals I've seen. Awesome job.

The only problem I see with it is that your onclick to expand a paper entry is not working under Chrome. Chrome and Safari both use WebKit, and WebKit does some event things a little weird; I'll see if I can fix it in such a way that it doesn't break under Chrome or Firefox, but someone else will have to test out IE and Opera.

How would you feel about adding a row of dropdowns with topic, subtopic, and a live text search box? (See http://ajaxpatterns.org/Live_Search for an explanation of what I mean.) I'd put it directly below the top navigation bar and above the "Articles" header, in the center column.
« Last Edit: March 19, 2010, 10:33:29 PM by badger »

no1uno

  • Global Moderator
  • Foundress Queen
  • *****
  • Posts: 681
Re: I can code for you guys
« Reply #38 on: March 20, 2010, 02:05:10 AM »
That (the use of dropdown menu's) is what I was talking about for the left-column of the new layout that I uploaded yesterday, I could, with a little fucking around, just put a couple of onclick events on the "tabs" (ie. change the links to tabs - link to the jscript file) and have all the code for all 5 "pages" in that one page...

It will be a little bulkier (maybe 2-3kb), but way lighter than most pages to load. Plus, everything on the userside is legit and above-board, anything we want to "hide" is on the server php (and obfuscated php ain't readable without a major fucking effort, the same can't be said of xhtml, css and jscript).

The articles we have may well approach 3K at the moment, but we could even hard code the dropdown menus, the Subject, Topic & Sub-topics that we are going to be interested in aren't that diverse. That saves a shitload of backend code and if you look around, the major issue remaining, the multi-level permissions, there are several articles on it at dreamincode, codeproject, etc.

That, plus the fact that a LOT of what we'd be linking to is already "public domain" and hosted on other sites (full-text), plus the fact that any djvu/pdf files that are uploaded are uploaded to ftp accts, not the db, will allow this to be clean and tidy and super-sleek.

Put the drop-downs in the left column of the various pages, a single DOM level object in the main column (which can then be rather easily modified, only one ID) and hardcode for the various "pages", that will make the backend a whole lot easier to design.

PS I was just thinking about it, save a lot of fucking around if I just made the 5 separate pages (Home, Subject Search, Author Search, Upload Content & Modify Content, I'll also make pages for Legal/Copyright and Contact for the bottom links) - keeping them clean-code (ie. no alternate), without major load-overhead (ie. they are tiny download and quick). I'll throw up a basic set of pages so you can see what I mean. They'll all utilise the same screen.css, so that will be easy enough.
« Last Edit: March 20, 2010, 03:40:57 AM by no1uno »
"...     "A little learning is a dang'rous thing;
    Drink deep, or taste not the Pierian spring:
    There shallow draughts intoxicate the brain,
    And drinking largely sobers us again.
..."

Naf1

  • Foundress Queen
  • *****
  • Posts: 753
Re: I can code for you guys
« Reply #39 on: March 21, 2010, 04:18:51 AM »
OK, so I have what we need!

http://www.orgchembase.com/refbase-0.9.5/

(before you start thinking thats great, but how do we get it? That is my site! Created for our use as a community!)
I have installed xoops with PubMedPDF on the root of that sever but I am thinking of changing to Typo3 or whatever badger recommends we push forward with.

Admin can add entries by simply adding the citation which can be downloaded from any of the sites hosting the journals (ingenta connect, sciencedirect, RSC and the many others) which makes it very easy then attach the pdf, you can search through the database but not download the files unless you are an account holder. It does not do fulltext search (yet) but can hold any type of file format Djvu included. The search function is very quick and works excellently, and advanced search even better you can search for DOI whatever you have, and it presents the information with a link to download the file for account holders, without link to the actual paper for non-account holders. Fulltext searching will be implemented in the backend by swish-e or similar in the near future and a fulltext search panel added into the php scripts. At the moment the only functionality disabled is the exporting of citations into certain formats (since the files will be there for immediate download its not really an issue), and I have yet to change the permissions of the actual pdfs so they are locked even to members at the mement. I just have to run a batch chmod so that will take 2 seconds! It stores for searching DOI, Title, Abstract, Keywords, publisher ect ect. I am adding papers now, I will add the rhodium/pdf and also /djvu papers first then move on from there. There are still some tweaks to do in the frontend and I would really appreciate any help anyone can give ;D As this is just one aspect of the site going together, I am hoping to provide online modeling tools a complete reference and book library and a lot of other smaller tools and apps that fit together to make one complete site.
« Last Edit: March 21, 2010, 04:26:36 AM by Naf1 »