Back to Silas S. Brown's home page

Web Access Gateway: Online Help

Welcome to the Web access gateway.  This is the online help file; it is not the same as the gateway's home page.

This gateway currently has two related purposes:

1.  To make the Web easier to access for the visually impaired;
2.  To allow speakers of other languages to view Web pages written in them, when the encodings are not supported by their browsers.

For help on a particular option, jump straight to the options list. There is also a common problems section.

Users of earlier versions may like to note that the language translation facility has been removed due to inadequacies; also the ``Leave smileys etc as they are'' option has been replaced by ``Perform smiley substitution'' (ie. substitution is now off by default), because it is slow for large Web pages and you should only enable it if you really want it.

A description of the gateway follows.

The Internet and the World Wide Web

Briefly, the Internet is a continuously changing global network of links between computers, arranged in such a way that any two computers connected to it may communicate with each other.  The World Wide Web is an arrangement that allows some of these computers to store documents, which may be viewed from other computers on the Internet.

Most of the documents on the World Wide Web are written in Hypertext Markup Language (HTML).  HTML allows links to other documents to be embedded within a document.  Should a link be selected, an attempt will be made to fetch and display the second document.  This may be stored on a completely different computer.  Hence, the whole thing can become a world wide ``web'' of documents and links between documents.

The documents that you view on the World Wide Web may not have been stored on any computer.  Instead, they may have been generated, by a computer program running on that computer, especially for your request.  As the user, you may know little difference between a program retrieving the document from storage and a program generating it.

Programs that generate documents can be given parameters.  This will usually involve you filling in some kind of ``form'' and ``submitting'' it.  The computer that your computer is communicating with will then be told what you have put in the form, and it can use this information while generating the document that it returns as a result (the remote computer can also store the information if necessary).

It is also possible to embed parameters in a link, to be put in a document.  Thus, if you as user were to activate the link, it would be as though you had filled in a form with those parameters.
 

``Hijacking'' the Web

Given the above, it is possible to ``hijack'' somebody's session on the World Wide Web.  Suppose that computer 1 wants to view a document stored on computer 2.  Rather than ask computer 2 for the document, suppose that it asks computer 3 instead.  The nature of the request includes exactly where the document is to be found, on computer 2.  A program is activated on computer 3, which gets the document from computer 2, and returns it to computer 1.  However, before returning it, it can change all of the links in the document to point to computer 3.  Therefore, when any of these links are followed, a similar process will happen, and so on - in effect, computer 3 has become a ``gateway'' to the World Wide Web for computer 1, arbitrating between computer 1 and the rest of the world.
 

Access Gateway

There may seem little point in the above arrangement, but the ``gateway'' computer, rather than just retrieving the documents that it is told to retrieve, can do something useful with them before it returns them.  For example, it can process some of the markup language in order to make the document easier to read for visually disabled people.  It is possible to configure exactly what is done, since no two people are alike (this is why it is not implemented as a proxy).

When you view the ``access gateway'', you have a form, which allows you to type in the URL of some other document on the World Wide Web, for the gateway to retrieve.  There are also several options that allow you to customise what the gateway does to pages.  To get to most of them, you need to press one of the buttons under the URL box; this arrangement is to keep the main form simple.

You do not have to set any of the options at this point; if for any reason you are not happy with them at some later point, or if you would like to jump to a different page on the World Wide Web that is not in any of the links, then you may return to the form later.  This is done by following a link at the top of whatever page is being displayed, called ``Change access options''.  When you do this, the options will be as you left them previously, except that the page's URL will be changed to whatever page you were just viewing, if it were different from the one that you had entered previously (because you had followed a link).

Saving your options: Once you have found your preferred options, you can save them on most browsers by either bookmarking the page, or setting it as your home or start page.  The method of doing this will depend on the software that you are using, so please consult the menus, the online help, the documentation, or the next person in the computer room.

Note that it is sometimes possible to accidentally leave the access gateway by following a link in a document, even though the program modifies links so that they ``go through'' the gateway.  This can happen, for example, if you have enabled Javascript, the document contains a Javascript program, and that program causes your Web browser to retrieve a different page directly.  Normally, this should not happen, although there are an increasing number of Javascript programs offering the user a menu of documents to retrieve.

In addition, the gateway can only process HTML documents.  If you retrieve a document that is not HTML, you will get a message about this, and a direct link to the document.

What follows are brief explanations of the various options on the form.  To submit the form, press the ``Get page'' button.  You can read on through the options or you can select a link.
 Disable all sight-related options
 Password button
 Characters button
 Access button
 Colours button
 Size button
 Options button
 Extensions button
 

Disable all sight-related options

If you have no sight problems (ie. you are using the access gateway for foreign languages etc), then select this box, which disables all of the sight-related options accessible by pressing the ``Access'' button.  You may then omit reading much of this document by clicking here.  Note: ``Enable scripts'' remains OFF when you select this option, because Javascript often includes code to manipulate URLs that is incompatible with Web mediators like the access gateway.

Options available under the ``Access'' button

You can read on through the options or you can select a link.
Scripts and applets
 Enable scripts
 If scripts not enabled, don't attempt to guess their links either
 Don't abbreviate URLs when displaying links
 Enable applets
 If applets not enabled, don't attempt to process their parameters either
 Enable embedded media
 Don't try to extract text from Flash
Frames and tables
 Leave frames as they are
 If leaving frames, make them generic
 Leave tables as they are
 If leaving tables, make them generic
Images
 Leave images in links as images
 Leave image maps as they are
 Repeat ALT tags after images
 Don't remove images from buttons
 Remove ALL images
Style
 Enable style sheets
 Use document body attributes
 Use document font settings
 Don't change blinking text to bold
 Don't change underlined text that is not a link to bold
 Don't change italic text to bold
 Don't change marquee text into plain text
Screenreader
 Don't attempt to remove surplus spaces between characters
 Don't move link banner to bottom
 Perform smiley substitution
 Don't put brackets around links
 Don't add ``End of web page''
 Enable delayed refresh/redirection
Miscellaneous
 Don't reformat long lines
 Don't strip HEIGHT with percentages
 Don't strip SPACER tags
 Don't expand Acronym tags
 Promote headings

Enable scripts

By default, all Javascript programs (and tag attributes beginning ``on'') are removed from the page that you are retrieving.  (``Noscript'' tags are also removed so any relevant messages are displayed.)  This is because some web sites add Javascript to open new browser windows with advertisements in them, or something similar, and these present problems for people using speech synthesisers or large fonts (some browsers behave very unusually when the text is too big to fit in the window size specified by the script, and the script often specifies that the window cannot be scrolled or resized).

If you really want the gateway to allow Javascript through, switch on this option.  Note, however, that the Javascript might manipulate URLs in a way that is incompatible with the gateway.  You might be taken out of the gateway, your settings might be overridden, or the page might not work at all.  Use with care.
 

If scripts not enabled, don't attempt to guess their links either

Normally, if scripts are disabled, then they are searched for anything that might be a URL, in order to cope with pages that use JavaScript as the sole means of navigation (many pages that are created by automatic page-generation software are like this).  This option disables that search.
 

Don't abbreviate URLs when displaying links

When the gateway guesses a link from Javascript, or otherwise has to output a link with no text (such as an image with no ALT), it displays the link's URL.  By default, the URL is abbreviated.  If you want to see the full URLs, select this option; the page is likely to be more cluttered.
 

Enable applets

By default, the program disables Java and other applets in the page.  Switch this option on to re-enable them.
 

If applets not enabled, don't attempt to process their parameters either

By default, if applets are not enabled, the program will attempt to find meaning in their parameters.  If it finds some apparently meaningful text, it will output it, and if it finds a URL it will generate a link.  Selecting this option will disable these attempts.
 

Enable embedded media

By default, the gateway will turn any EMBED elements (for music, `flash' movies, etc) into links, so that the user is in more control over when (or if) they run; this is useful if your browser tends to crash when it encounters this media.  Selecting this option will allow the EMBED elements through; it also implies ``Don't try to extract text from Flash'' (below).
 

Don't try to extract text from Flash

By default, if the gateway is turning EMBED elements (for music, `flash' movies, etc) into links, and the link points to a .swf (Flash) file, then the gateway will retrieve it and try to extract any text and links from it.  This is because some sites put their main navigation into Flash files, and extracting the links from them is the only way to navigate the site if you cannot use ShockWave.  Selecting this option will prevent the gateway from trying to interpret Flash files.  The ``Enable embedded media'' option implies this option.
 

Leave frames as they are

``Frames'' are an arrangement whereby a document can be made up of several other documents.  They are frequently a problem for blind users using speech synthesisers, and also partially sighted users who have set large fonts, especially if the page author did not think of this possibility while specifying the frame parameters.  By default, the program removes all frames.  Instead, it retrieves the documents that they contain, and inserts them into the document, one after another, separated by horizontal rules.  If you would not like this, then switching this option on will cause the program to leave frames as they are.
 

If leaving frames, make them generic

This is only effective if ``Leave frames as they are'' is also switched on.  Some people are able to read frames, but may struggle if they force a particular formatting, such as preventing scrolling and resizing.  The ``leave frames as they are'' option will not affect such formatting, but selecting both it and this will cause the formatting to be removed.

You may like to select this but leave the ``Leave frames as they are'' option off.  In this case, frames will be written out, but, should you find a frame that really would be easier as a frame, you can switch on ``Leave frames as they are'' and the formatting will still be removed.
 

Leave tables as they are

Tables are often incorrectly read by speech software, especially if they are used to format text into multiple columns on the page.  What often happens is that a line is read from each column in turn.  Tables can also be a problem to people using large fonts, if the browser cannot get the table to fit on the screen.  By default, the access gateway simply writes out tables into the document, row by row, with colon separators between the columns and horizontal lines betwen the rows (although there are some precautions against excessive numbers of horizontal lines being drawn when a table is there purely for layout purposes).  If you would prefer tables to be left as they were, then switch on this option.

Please note that tables may still be corrupted by the code that moves link banners to the bottom.  If you select this option, then you might like to consider also selecting ``don't move link banner to bottom'' below.
 

If leaving tables, make them generic

This is only effective if ``Leave tables as they are'' is also switched on.  Some people are able to read tables, but may struggle if they force a particular formatting, font, colour, or background.  The ``leave tables as they are'' option will not affect such formatting, but selecting both it and this will cause the formatting to be removed.  It is still possible for tables to be larger than the screen width.

You may like to select this but leave the ``Leave tables as they are'' option off.  In this case, tables will be written out, but, should you find a table that really would be easier as a table, you can switch on ``Leave tables as they are'' and the formatting will still be removed.
 

Leave images in links as images

A link can be an image as well as a piece of text.  However, blind and partially sighted people rarely realise that a given image is a link, and could thus miss a link, especially if an entire contents page were formed of these.  By default, the program replaces images in links with their ``alternative'' text.  This is text that is hidden in the image specification, designed for Web browsers that cannot support images.  Unfortunately, some images do not have alternative text; the gateway tries to help with these by extrapolating from the linked-to URL, but sometimes even that is not meaningful. If you would like to see the images that are in the links, then switch on this option.
 

Leave image maps as they are

By default, the program replaces client-side image maps with text links, so that they can be navigated by people who are unable to use the image.  If you would prefer image maps to be left as they are, then switch on this option.
 

Repeat ALT tags after images

Causes all ALT tags to be repeated after images as well as included in them.  This is useful for some graphical browsers that do not display ALT text even when you switch off images, or if you are partially sighted and want to see the non-textual images but read the text in your own font.  The option is not checked by default because it will cause unnecessary repetition if you are using a speech synthesiser and an accessible browser.
 

Don't remove images from buttons

By default, the gateway removes all images from form submit buttons that it encounters.  Some browsers allow buttons to be displayed at any size and/or spoken, but not when the button is an image.  For this reason, the gateway removes these images.  Select this if you want to see the images on the buttons.
 

Remove ALL images

This option will cause all images to be removed, including their ALT tags.  If you want to see the ALT tags, also select ``Repeat ALT tags after images'' above.
 

Enable style sheets

By default, anything in a STYLE tag is disabled, and so are LINK, DIV and SPAN tags plus any STYLE attributes.  Selecting this will enable them again.  This can allow documents to override your options by using cascading style sheets.
 

Use document body attributes (background etc)

By default, the program will remove any general text and background colour that is set, as well as any background image that is set.  This means that the user may specify their own colours, which are easier to read.  As well as this, some page authors set a text colour in such a way that browsers that are told to ignore the background colour will still pay attention to the text colour.  If the text colour is the same as the background colour, or if there is little difference, then it may be impossible to read the page without pasting it into a wordprocessor or doing something similar.  If you would like to see the document's original colours and background images, switch this option on.
 

Use document font settings

By default, any attempts that the document makes to change the font size, type, or colour, are ignored.  If you would not like them ignored, switch this option on.  Note that this does not cover things such as bold, italic and underline, which are dealt with separately.
 

Don't change blinking text to bold

By default, the program changes blinking text into bold text, because blinking text can be difficult to read.  If you would not like this, switch this option on.
 

Don't change underlined text that is not a link to bold

By default, the program changes underlined text into bold text, because a link to another document is often represented by a combination of underlining the text and displaying it in a different colour.  If a user is colourblind, or if the colour of the text happens to match the user's link colour, then underlined text can seem like a link, which can be confusing.  If you would prefer underlined text left as it is, then switch on this option.
 

Don't change italic text to bold

By default, italic text is changed into bold text, because italic text can be difficult to read, especially if there is lots of it.  If you would prefer italic text left as it is, then switch on this option.
 

Don't change marquee text into plain text

By default, marquee text, which scrolls horizontally across the screen, is changed into plain text, so that the user can read it at their own pace, and so that speech software does not get confused.  If you would prefer to leave marquee text as it is, then turn on this option.  Note that some pages use Java applets rather than the marquee tag, and this option does not affect these.
 

Don't attempt to remove surplus spaces between characters

By default, the gateway will look for places where HTML authors have placed a space between each character, as in ``i n t r o d u c t i o n''.  Although this may be considered by some to look good, it can cause word-wrap problems with large fonts and causes most speech synthesisers to spell everything out.  The gateway attempts to remove these spaces.  However, it requires some guesswork to distinguish between these spaces and legitimate single-letter words in any language, and in some unusual cases it can go wrong.  Selecting this option disables the gateway's attempts to remove these spaces.
 

Don't move link banner to bottom

Web pages are increasingly putting lots of links and advertisements at the top, so it is difficult to get to the actual contents of the page if you have a sight problem.  By default, the gateway will try to detect these banners and move them to the bottom.  Select this option to suppress this.

The algorithm used to detect banners is liable to change; from time to time I try to improve it.  Ideally it should move the banner and no more than the banner (it should not move the start of the text to the bottom); it should not interfere with small, manageable banners and it should not interfere with genuine index or contents pages.  However, deciding what general guidelines to give to a computer that doesn't understand the content of the page is not easy to do well.
 

Perform smiley substitution

On the Internet in general, it is fairly common for certain sequences of punctuation to be used to indicate facial expressions.  For example, a colon, a hyphen, and a closing parenthesis indicate a grin.  This is because, in printed form, a colon is like two small solid circles, one directly above the other; a hyphen is like a short horizontal line at a height about half way between the two circles, and a closing parenthesis is like a part of the right hand side of a circle.  If you rotate the whole thing by a quarter-turn clockwise, you get the two small solid circles aligned horizontally rather than vertically, then the short line underneath them at approximately their midpoint, now vertical, and then the part of a circle, now taken from the bottom of one.  With a bit of imagination, these look roughly like a pair of eyes, a nose, and a grinning mouth.

If you enable this option, the program will replace this idiom with the word ``grin'' in parentheses, so that it can be read properly by a speech synthesiser.  It also replaces numerous other similar idioms.  However, switching on this option also slows down the program considerably for large Web pages, so you should only use it if you need to.
 

Don't put [ ] around links

By default, opening and closing square bracket characters are placed around all links that don't already have them, to make them more easily distinguishable to blind people using speech synthesisers.  If you would prefer links left as they were, then switch on this option.
 

Don't add ``End of web page.''

This suppresses the addition of the text ``End of web page.'' at the end of a page, which can be useful for blind people or to confirm that a page is blank (or completely inaccessible even to the gateway).
 

Enable delayed refresh/redirection

By default, automatic page refreshes are disabled, and instead some text is printed with a link.  This is so that users can refresh pages at their own speed.  Selecting this option will cause the page refreshes to be done automatically.  However, if the gateway is processing frames, then frame refreshes will not be done automatically regardless of the setting of this option.
 

Don't reformat long lines

Long lines can cause problems for people using large print because some browsers format the entire page to the width of the longest line, even if it is wider than the screen.  This causes the page to require excessive horizontal scrolling.  Some of the causes of long lines are NOBR (no line break) tags, non-breaking spaces, and excessively-long words (such as URLs).  By default, the gateway takes measures to reduce this problem while not compromising on readability.  If you prefer that the gateway did not do this, select this option.
 

Don't strip HEIGHT with percentages

Some images that are used in frames are given tags like HEIGHT="100%", to cause them to fill a small frame.  When frames are removed, or the page is otherwise reformatted, these images can hinder page display (by, for example, causing one screenful of the page to be completely black), since they are usually not real images but only layout aids.  By default, the access gateway removes any HEIGHT attribute it finds with a percentage specification in it, in order to resolve this problem.  Select this option to suppress this behaviour.
 

Don't strip SPACER tags

SPACER tags are sometimes used for layout, but, when frames and tables are removed, the remaining SPACER tags can result in excessive amounts of space on the page.  By default SPACER tags are removed; use this option to stop that.

Don't expand Acronym tags

HTML 4.0 defines an ``acronym'' tag, which allows an acronym's expansion to be hidden in the HTML.  Many browsers do not display these, so the gateway expands them.  Use this option to prevent such expansion.

Promote headings that are below...

Some browsers display deeply-nested subheadings (eg. headings six levels deep) in a very small font, regardless of the font settings.  This option allows you to have such headings ``promoted'', ie. raised in level to the value you give.  By default, all headings below the third level deep are raised to that level.  In order to keep the nesting information, though, a hyphen and a space is prepended to the heading for each level of promotion.

Password

The password options are for pages that require authentication; most pages do not, and you can leave these blank unless you know that you will need a password.  The password is not encrypted in any way when it is sent to the gateway, and, if you are very concerned about security, then you should not use the gateway.

Options available under the ``Options'' button

You can read on through the options, or select one of these links:
Search
 Delete table rows not matching query
 Treat as frames links that match query
Navigation
 URL box at top of document
 Make URL box longer
 No change access options link
 No character set controls
 Don't show date stamp
 Don't add status line code to links
Browser quirks
 Avoid switching to/from SSL
 Replace audio EMBED with BGSOUND
 Fix Netscape 4.0 font/link bug
Cookies
 Use cookies for the options
 Don't store remote session IDs (only available on some servers)
Miscellaneous
 Replace paragraphs with indentation
 Make all text areas word wrap
 Remove document head
 Make reset buttons say reset
 Show hidden form fields
 Remove ALT tags (for non-readers)
 Replace Windows quotes etc

Delete table rows not matching query

Typing a query here causes all table rows not matching that query to be deleted.  This can be useful for those pages that present large amounts of information in a tabular form.

A query is either a list of words separated by spaces (in which case all of the words must be present for the query to match), or a set of alternative lists separated by commas.  For example, ``research groups, PhD'' will match text containing both ``research'' and ``groups'', and also text containing ``PhD''.  Case does not matter.  Note that queries do not remain in effect when you follow links.
 

Treat as frames links matching query

Typing a query here causes all links whose text matches the query to be treated as though they were frames, and the documents that they link to will be expanded into the document.  This implies that frame processing will be turned on, ie. frames will be expanded into the document, even if you have disabled the sight-related access options.  To avoid excessive crawling, there is a depth limit on the number of documents that can be retrieved at once (as there is with frames), and no URL will be retrieved twice.  This option is intended for occasional use and offline browsing; you cannot have a query permanently set while you browse, as server load must be considered.  Note that the pages themselves will not be retrieved and searched; only the text within the link (ie. between the <A> and the </A>) will be searched.

A query is either a list of words separated by spaces (in which case all of the words must be present for the query to match), or a set of alternative lists separated by commas.  For example, ``research groups, PhD'' will match text containing both ``research'' and ``groups'', and also text containing ``PhD''.  Case does not matter.  Note that queries do not remain in effect when you follow links.
 

URL box at top of document

Selecting this option causes an input box to appear before the ``change access options'' link on each page, containing the page's URL.  Changing this URL and pressing Enter will retrieve the new page (using the current options, of course).  Please note that some old browsers have trouble with form submission if there is more than one form on a page.  If you are using such a browser, then you may find that pages with forms are unusable with this option on; in this case, follow the ``Change access options'' link and switch it off.
 

Make URL box longer

This causes the URL box to be longer.
 

No ``change access options'' link

This will suppress the display of the ``change access options'' link at the top of pages.  Please be careful that you know what you are doing when you select this, since it will prevent you from changing the options in future (unless you manually edit the CGI parameters to get them back).  Its normal use is when you know that a certain group of people using the gateway are likely to always want the same options and don't want to worry about setting them.  See also ``Don't show date stamp'' (below).
 

No character set controls

When displaying characters from other languages, the gateway normally inserts a small control that shows which character set it has detected and allows you to override this.  This option will suppress that control.  Please be careful that you know what you are doing when you select this, since it will prevent you from changing the character set (unless you manually edit the CGI parameters).  It might be useful when you want to print out a page.
 

Don't show date stamp

By default, the page's ``Last-Modified'' date is added at the top (when it is available); this option will remove it.
 

Don't add status line code to links

Many graphical browsers show the URL of a link in the status line, when the mouse is placed over that link.  This is not usually helpful with gateway URLs, so by default the gateway outputs Javascript that makes the status line indicate the URL that the gateway would retrieve, rather than the full gateway query.  However, in a document with lots of links, this Javascript can significantly increase the size of the document, which is a bad thing for low-bandwidth devices such as PDAs and Internet mobile phones (which probably don't have status lines anyway).  The Javascript can be suppressed by this option.  The gateway will try to select this option automatically if your browser does not support Javascript.
 

Avoid switching to/from SSL

This option appears on gateways that are capable of switching in/out of SSL mode, both server-side and client-side.  Normally, if the gateway has this capability, it will encrypt its connection to your browser when (and only when) you are accessing `secure' pages (although this behaviour is based on following links from within the gateway; if you were to type a URL then your browser will use a secure or insecure connection depending on what your connection is at the moment).  Note that this is about the connection to your BROWSER, not the connection to the remote site (the gateway's ability to use encrypted connections to the remote site is independent of its ability to use encrypted connections to your browser).

If your browser does not support SSL (encrypted pages), you should check this option if it appears, and note that you have to trust your link to the gateway when you are using it for secure sites.  Also if you are using an SSL gateway and you want it to use ONLY encrypted connections to your browser (regardless of whether or not the remote site is SSL) then you should use this option to stop the gateway from switching to non-SSL.
 

Replace audio EMBED with BGSOUND (for old versions of Internet Explorer)

If you are running an old version of Internet Explorer and you switch this option on, then you should be able to hear background music more often.
 

Fix Netscape 4.0 font/link bug

Netscape Communicator 4.05 (and possibly some other versions) has a bug in that, if a document changes style within a link and fails to change back before the end of the link, the rest of the document (or table) will look like a link, thus making it very confusing.  Many Web authors fail to notice this in cases where the end of a table prevents it from happening, which means that it will occur when the access gateway removes all tables.

It would not be feasible for the access gateway to cope with all possible changes of style within a link and re-write the HTML accordingly, since some style changes may be introduced after the access gateway has been written.  Further, future versions of Netscape may correct the problem.  This box, then, is a ``quick hack''.  Selecting it causes the font to be reset at the end of every link.  A side effect of this is that any changes that documents make to the font may be reversed in places.  This does not matter if you are overriding document fonts anyway.
 

Remove document head

Some documents have browser-specific HTML in their HEAD sections, causing some browsers to refuse to load them (whether or not they are through the access gateway).  If you come across such a document, try checking this option, which removes the HEAD section of the document altogether (although the gateway still adds its own HEAD section).
 

Use cookies for the options

The gateway cannot store the options you set on the server.  Normally, it adds details of them to every link in every Web page, so that it can use them if you follow that link.  This makes some pages unnecessarily large and slow to download on low-bandwidth links.  Some browsers support ``cookies'', which allow the gateway to store the options in your browser and have them sent back no matter which link you follow, alleviating the need to add them to every link.  However, cookies can be annoying for people who have set their browsers to warn about them, so the gateway will not attempt to use them unless this option is turned on.

If this option is selected, then the gateway will send cookies, but it won't start removing the redundant options from the links until it has confirmed that your browser is returning its cookies.  This normally happens on the second page that you fetch, but you might like to check that cookie support is turned on in your browser.
 

Don't store remote session IDs

(only available on some servers)

Many websites use ``cookies'', which are small amounts of data that are stored by your browser and sent back to the site every time you get a new page.  They are usually used to identify your session (like a customer number) so that the site can collect statistics about their visitors.

Cookies do not work through mediators like the access gateway, but some sites are made in such a way that they don't work without them (and sometimes it's not obvious why they're not working).  For this reason, the gateway will store any cookies the site sends as settings in your options, to be sent back to the site.  The space for storing cookies is limited; it is not intended to be a full implementation of the cookies standard, only enough to allow you access to sites that refuse entry without cookies.

If you prefer that the gateway ignores cookies sent by remote sites, use this option.

In any event, only cookies from documents are stored; cookies from images are dealt with by your browser, since images are fetched directly and not through the gateway.
 

Replace paragraphs with indentation

Causes all paragraph breaks to be replaced with line breaks followed by indentation.  Some people prefer this to the blank line that is left by most browsers.
 

Make all text areas word wrap

This causes the gateway to add the WRAP=HARD attribute to any TEXTAREA tag it finds (replacing any previous WRAP attribute).  This means that text will automatically wrap as you type into text areas on forms, even if the form designer did not intend this.  Some people prefer this.  Use this option with care as some forms will require unwrapped text (eg. if you have to paste code into the box).
 

Make reset buttons say ``Reset''

Documents are sometimes very unclear about what the buttons on their forms actually do.  Enabling this option causes any reset buttons to say ``Reset'', rather than whatever was in the document.  Submit buttons are not changed because they can sometimes carry extra information that would be lost, but knowing which button is Reset is sometimes helpful.
 

Show hidden form fields

This causes any HIDDEN inputs in forms the server sends to be converted into TEXT inputs, so that you can modify them.  Sometimes useful when you want to access servers that don't let you in without Javascript (since they sometimes implement the test by getting the Javascript to change the value of a hidden form option).
 

Remove ALT tags (for non-readers)

Non-readers will probably not be reading this, but their instructors might.  ALT tags have a tendency to pop up on recent browsers, and this can be distracting.  This option will remove all ALT tags.  Note: Make sure ``Disable all sight-related options'' is also set when you set this.
 

Replace Windows quotes etc

The Windows-1252 character set adds some extra characters to ISO-Latin-1, such as `smart' quotes.  These are not often displayed on non-Microsoft systems.  Selecting this option will cause the gateway to substitute other characters for some of them.  It is only effective in languages covered by Latin-1.

Colours

These options allow you to change the colour scheme of a page.  They may not work if you have specified that the document's own colours should not be ignored (or you have disabled all sight-related options).  ``Use browser setting'' is the default in all cases.  If you are an expert in HTML programming, then you can modify the RGB values directly by changing the CGI in the request to any values that you like.  However, it would be impractical to put nearly 17 million colour options on the form, especially given that some browsers display only a limited number of them.  Therefore, only a few options are available.

Font size

This allows you to change the size of text displayed on pages.  If you have told the program not to ignore the document's own fonts (or have disabled all sight-related options), then these may override any size that you specify here.  The default is to use whatever size your browser is configured to use.  Only text is affected, not images or text in images.

Some browsers (notably NCSA Mosaic) do not support different font sizes, but do display their headings larger.  If you are stuck with one of those then you can choose one of the headings instead of a font size.  However, due to a limitation in HTML, this has side-effects, notably the addition of many line breaks (especially if you have the ``fix Netscape font/link bug'' option turned on, which you probably don't need if you don't have Netscape).


Extensions

If this button is present, then it allows you to enable any third-party extensions that may have been added to the copy of the gateway on this server.

Other Languages and Encodings

If you see nonsense characters when you try to view Web pages written in a different alphabet (or with ideographic characters), and you cannot configure your browser, then you may find the conversion facility useful.  Here is a simple explanation of the problem:

Computers work in numbers, and a Web page is really lots of numbers.  Every letter of the English alphabet has a number, and your browser knows which English letter or symbol to display for each number.  Other alphabets also have numbers, and things like Chinese characters have big numbers because there are lots of them.  The big numbers have to be split into smaller numbers, because each number has to fit in a small space.  Numbers for other alphabets and for Chinese characters are often the same numbers as the ones that are used by English, so, if the computer is programmed to display English, then pages in other languages might look like random English letters and symbols.

Even if the computer is programmed to display another language, it can still go wrong sometimes.  This is because some languages have several different ways of numbering the letters, or different ways of splitting the big numbers into smaller ones, and if your computer is programmed to use one or more of these ways then you can still get pages that are written in other ways.

This program can read pages that are written in one of these ways, and give them to your computer as pictures.  Pages are not normally written in this way, partly because it takes more time and computer memory.  The gateway can also be asked to give the characters to your computer in other encodings (such as UTF-8); this is quicker than pictures but is not supported by all browsers.

You normally only need to tell the program which language the page is in.  The languages in the drop-down boxes are sorted alphabetically by their English names, and you can get near a language by pressing its first letter on most browsers.  The program will detect between the various types of encodings appropriate to the language, and I hope to improve this detection code in future.  If you want to override the detection to a particular encoding, then you can do this by pressing ``Override'' after getting the page.

If you have some experience in converting between encodings, then you may be concerned about documents that encode high bytes as HTML numerical ampersand sequences.  This program should cope with them automatically.  If for some reason you do not want this to happen, select the ``Don't decode HTML escapes before interpreting characters'' box.

Normally the gateway will try to detect the character set being used by the text, independently of what the Web server says it is, since many Web servers are configured to give wrong information.  However, the gateway does process certain ``charset'' headers, if they are appropriate to the language being displayed.  Select the ``Always ignore `charset' headers'' box if you would prefer the gateway to always ignore these headers.

The ``Images (ask browser to enlarge)'' option, if present, will cause the HEIGHT and WIDTH tags to be doubled in value.  On some web browsers (including recent versions of Netscape and Internet Explorer), this causes the characters to be doubled in size.

The ``alternative base URL for images'' box is mainly for situations where more than one image server is available.  It is normally set to a sensible default by the webmaster on the language entry form, and can be left alone unless you really know what you're doing.

Note: The access gateway program is written in such a way that it can be run on different web servers.  Some webmasters may decide not to include the encoding options, or not to include the images, to save space.  If you find that these options are not available, then that is the explanation.


Common Problems

  1. The web server tells me to go away and come back with a Version 4 browser or cookies.
    This should not normally happen, since the access gateway pretends to be a Version 4 browser and emulates cookies if it can.  However, some installations have a dated version of Lynx (the program the gateway uses to get the web pages), and versions of Lynx before 2.8 do not allow browser fakery.  Try a different installation if possible.
  2. When I try to view a language, all the characters come out as squares.
    Check that the language is supported by the image server you are using.  Check that nobody has set a wrong URL under the Characters button (try deleting any URL there).
  3. Some of the gateways are out of date
    Badger their webmasters.  I have no control over the ones that are out of date, although I do plea that they keep in step.
  4. Why do I always get Japanese-style characters while I am not Japanese?
    Some Traditional or Simplified Chinese characters, and some Korean characters, use the same unicode codepoints as Japanese.  The gateway should use the appropriate versions for Chinese and Korean pages, but if the correct versions are not available on the server you are using then the gateway might attempt to use the Japanese versions instead.  If this proves too difficult to read, try using a different server.
  5. The language options are not available
    This is because not all servers have installed them.  Some webmasters are concerned about the load on their servers.
  6. The gateway still doesn't make the page accessible.
    Perhaps it relies on images a lot, which contain textual information and do not have ALT tags, or perhaps it relies on plug-ins.  There is little I can do about that short of OCR.  Write to the webmaster.

All material © Silas S. Brown unless otherwise stated.