One option would be to undertake the dualling along the existing alignment of the route. Potential advantages of this option include the directness of the route, and the limitation of environmental impacts, as the route would be within the existing transport corridor. However, potential disadvantages include the increased difficulty of undertaking the construction “on-line” with associated additional traffic management costs, disruption, delays, and increased safety risks.
http://je.pl/drrk

After the publication in July 1972 of the Urban Motorways Committee report "New Roads in Towns" the proposals were reviewed. That review led to three "on-line" and two western bypass alternatives being identified as possible improvements between Roestock and Stanborough.
http://www.iht.org/motorway/hatfield.htm

Both on-line, and off-line widenings to the east and west of the existing motorway, were investigated.
http://je.pl/x6ef

A slightly neglected footbridge over the bypass: this shows, if the alignment didn't already prove it, that there's no chance of an on-line upgrade to three lane motorway here.
http://www.cbrd.co.uk/media/photo/gallery11.shtml [drugi obrazek]

An initial review was carried out to determine sections of the existing N9 which were within the settlement areas and would require an off-line bypass. There were proposals for bypasses at Carlow, Castledermot, Thamestown and Lukeswell. Secondly there are lengths of the N9 where the horizontal alignment is below standard of a high quality dual carriageway / motorway. For these sections there is no alternative but to go "off-line" to achieve the necessary standard. Examples of this arrangement are shown in RSR2 Figure 6.6 and RSR3 Figures 6.6, 6.7 and 6.8. These factors reduced the length of possible on-line improvement from 107km to 50km (47% of total length). Over 53% of the route would need to be "off-line" to avoid settlements and to improve the horizontal alignment.

A sample length of on-line route was studied in more detail and a number of major disadvantages were identified:
(i) Service roads would be needed over significant lengths of the on-line section to maintain access to properties and land. This would increase the cross-section, verge to verge, compared to an off-line solution, from 26 metres to 49 metres.
(ii) A significant number of properties close to the existing N9 would have to be demolished to accommodate this very wide cross-section.
(iii) Access for communities located in the vicinity of the N9 would be severely affected by restricting access to the new road. Local journeys and journeys to and from the new road would be forced to use low standard rural roads.
(iv) Severance would be significant for properties and land currently fronting the N9.
(v) Extensive traffic management measures would be required during construction which would increase the number of accidents and increase construction costs.
(vi) There would be significant delays to traffic on the existing N9 during construction.
For the above reasons the "on-line" option was discarded.
http://www.n9-n10kilcullen-waterford.ie/route-selection/2-5N...

budowa + w śladzie + droga
http://je.pl/n4nb